title
stringlengths 5
246
| categories
stringlengths 5
94
⌀ | abstract
stringlengths 54
5.03k
| authors
stringlengths 0
6.72k
| doi
stringlengths 12
54
⌀ | id
stringlengths 6
10
⌀ | year
float64 2.02k
2.02k
⌀ | venue
stringclasses 13
values |
|---|---|---|---|---|---|---|---|
Improving Neural Relation Extraction with Positive and Unlabeled Learning
| null |
We present a novel approach to improve the performance of distant supervision relation extraction with Positive and Unlabeled (PU) Learning. This approach first applies reinforcement learning to decide whether a sentence is positive to a given relation, and then positive and unlabeled bags are constructed. In contrast to most previous studies, which mainly use selected positive instances only, we make full use of unlabeled instances and propose two new representations for positive and unlabeled bags. These two representations are then combined in an appropriate way to make bag-level prediction. Experimental results on a widely used real-world dataset demonstrate that this new approach indeed achieves significant and consistent improvements as compared to several competitive baselines.
|
Zhengqiu He, Wenliang Chen, Yuyi Wang, Wei Zhang, Guanchun Wang, Min Zhang
| null | null | 2,020 |
aaai
|
Leveraging Multi-Token Entities in Document-Level Named Entity Recognition
| null |
Most state-of-the-art named entity recognition systems are designed to process each sentence within a document independently. These systems are easy to confuse entity types when the context information in a sentence is not sufficient enough. To utilize the context information within the whole document, most document-level work let neural networks on their own to learn the relation across sentences, which is not intuitive enough for us humans. In this paper, we divide entities to multi-token entities that contain multiple tokens and single-token entities that are composed of a single token. We propose that the context information of multi-token entities should be more reliable in document-level NER for news articles. We design a fusion attention mechanism which not only learns the semantic relevance between occurrences of the same token, but also focuses more on occurrences belonging to multi-tokens entities. To identify multi-token entities, we design an auxiliary task namely ‘Multi-token Entity Classification’ and perform this task simultaneously with document-level NER. This auxiliary task is simplified from NER and doesn't require extra annotation. Experimental results on the CoNLL-2003 dataset and OntoNotesnbm dataset show that our model outperforms state-of-the-art sentence-level and document-level NER methods.
|
Anwen Hu, Zhicheng Dou, Jian-Yun Nie, Ji-Rong Wen
| null | null | 2,020 |
aaai
|
Knowledge-Enriched Visual Storytelling
| null |
Stories are diverse and highly personalized, resulting in a large possible output space for story generation. Existing end-to-end approaches produce monotonous stories because they are limited to the vocabulary and knowledge in a single training dataset. This paper introduces KG-Story, a three-stage framework that allows the story generation model to take advantage of external Knowledge Graphs to produce interesting stories. KG-Story distills a set of representative words from the input prompts, enriches the word set by using external knowledge graphs, and finally generates stories based on the enriched word set. This distill-enrich-generate framework allows the use of external resources not only for the enrichment phase, but also for the distillation and generation phases. In this paper, we show the superiority of KG-Story for visual storytelling, where the input prompt is a sequence of five photos and the output is a short story. Per the human ranking evaluation, stories generated by KG-Story are on average ranked better than that of the state-of-the-art systems. Our code and output stories are available at https://github.com/zychen423/KE-VIST.
|
Chao-Chun Hsu, Zi-Yuan Chen, Chi-Yang Hsu, Chih-Chia Li, Tzu-Yuan Lin, Ting-Hao Huang, Lun-Wei Ku
| null | null | 2,020 |
aaai
|
ManyModalQA: Modality Disambiguation and QA over Diverse Inputs
| null |
We present a new multimodal question answering challenge, ManyModalQA, in which an agent must answer a question by considering three distinct modalities: text, images, and tables. We collect our data by scraping Wikipedia and then utilize crowdsourcing to collect question-answer pairs. Our questions are ambiguous, in that the modality that contains the answer is not easily determined based solely upon the question. To demonstrate this ambiguity, we construct a modality selector (or disambiguator) network, and this model gets substantially lower accuracy on our challenge set, compared to existing datasets, indicating that our questions are more ambiguous. By analyzing this model, we investigate which words in the question are indicative of the modality. Next, we construct a simple baseline ManyModalQA model, which, based on the prediction from the modality selector, fires a corresponding pre-trained state-of-the-art unimodal QA model. We focus on providing the community with a new manymodal evaluation set and only provide a fine-tuning set, with the expectation that existing datasets and approaches will be transferred for most of the training, to encourage low-resource generalization without large, monolithic training sets for each new task. There is a significant gap between our baseline models and human performance; therefore, we hope that this challenge encourages research in end-to-end modality disambiguation and multimodal QA models, as well as transfer learning.
|
Darryl Hannan, Akshay Jain, Mohit Bansal
| null | null | 2,020 |
aaai
|
Privacy Enhanced Multimodal Neural Representations for Emotion Recognition
| null |
Many mobile applications and virtual conversational agents now aim to recognize and adapt to emotions. To enable this, data are transmitted from users' devices and stored on central servers. Yet, these data contain sensitive information that could be used by mobile applications without user's consent or, maliciously, by an eavesdropping adversary. In this work, we show how multimodal representations trained for a primary task, here emotion recognition, can unintentionally leak demographic information, which could override a selected opt-out option by the user. We analyze how this leakage differs in representations obtained from textual, acoustic, and multimodal data. We use an adversarial learning paradigm to unlearn the private information present in a representation and investigate the effect of varying the strength of the adversarial component on the primary task and on the privacy metric, defined here as the inability of an attacker to predict specific demographic information. We evaluate this paradigm on multiple datasets and show that we can improve the privacy metric while not significantly impacting the performance on the primary task. To the best of our knowledge, this is the first work to analyze how the privacy metric differs across modalities and how multiple privacy concerns can be tackled while still maintaining performance on emotion recognition.
|
Mimansa Jaiswal, Emily Mower Provost
| null | null | 2,020 |
aaai
|
Discontinuous Constituent Parsing with Pointer Networks
| null |
One of the most complex syntactic representations used in computational linguistics and NLP are discontinuous constituent trees, crucial for representing all grammatical phenomena of languages such as German. Recent advances in dependency parsing have shown that Pointer Networks excel in efficiently parsing syntactic relations between words in a sentence. This kind of sequence-to-sequence models achieve outstanding accuracies in building non-projective dependency trees, but its potential has not been proved yet on a more difficult task. We propose a novel neural network architecture that, by means of Pointer Networks, is able to generate the most accurate discontinuous constituent representations to date, even without the need of Part-of-Speech tagging information. To do so, we internally model discontinuous constituent structures as augmented non-projective dependency structures. The proposed approach achieves state-of-the-art results on the two widely-used NEGRA and TIGER benchmarks, outperforming previous work by a wide margin.
|
Daniel Fernández-González, Carlos Gómez-Rodríguez
| null | null | 2,020 |
aaai
|
Latent Relation Language Models
| null |
In this paper, we propose Latent Relation Language Models (LRLMs), a class of language models that parameterizes the joint distribution over the words in a document and the entities that occur therein via knowledge graph relations. This model has a number of attractive properties: it not only improves language modeling performance, but is also able to annotate the posterior probability of entity spans for a given text through relations. Experiments demonstrate empirical improvements over both word-based language models and a previous approach that incorporates knowledge graph information. Qualitative analysis further demonstrates the proposed model's ability to learn to predict appropriate relations in context. 1
|
Hiroaki Hayashi, Zecong Hu, Chenyan Xiong, Graham Neubig
| null | null | 2,020 |
aaai
|
Corpus Wide Argument Mining—A Working Solution
| null |
One of the main tasks in argument mining is the retrieval of argumentative content pertaining to a given topic. Most previous work addressed this task by retrieving a relatively small number of relevant documents as the initial source for such content. This line of research yielded moderate success, which is of limited use in a real-world system. Furthermore, for such a system to yield a comprehensive set of relevant arguments, over a wide range of topics, it requires leveraging a large and diverse corpus in an appropriate manner. Here we present a first end-to-end high-precision, corpus-wide argument mining system. This is made possible by combining sentence-level queries over an appropriate indexing of a very large corpus of newspaper articles, with an iterative annotation scheme. This scheme addresses the inherent label bias in the data and pinpoints the regions of the sample space whose manual labeling is required to obtain high-precision among top-ranked candidates.
|
Liat Ein-Dor, Eyal Shnarch, Lena Dankin, Alon Halfon, Benjamin Sznajder, Ariel Gera, Carlos Alzate, Martin Gleize, Leshem Choshen, Yufang Hou, Yonatan Bilu, Ranit Aharonov, Noam Slonim
| null | null | 2,020 |
aaai
|
ABSent: Cross-Lingual Sentence Representation Mapping with Bidirectional GANs
| null |
A number of cross-lingual transfer learning approaches based on neural networks have been proposed for the case when large amounts of parallel text are at our disposal. However, in many real-world settings, the size of parallel annotated training data is restricted. Additionally, prior cross-lingual mapping research has mainly focused on the word level. This raises the question of whether such techniques can also be applied to effortlessly obtain cross-lingually aligned sentence representations. To this end, we propose an Adversarial Bi-directional Sentence Embedding Mapping (ABSent) framework, which learns mappings of cross-lingual sentence representations from limited quantities of parallel data. The experiments show that our method outperforms several technically more powerful approaches, especially under challenging low-resource circumstances. The source code is available from https://github.com/zuohuif/ABSent along with relevant datasets.
|
Zuohui Fu, Yikun Xian, Shijie Geng, Yingqiang Ge, Yuting Wang, Xin Dong, Guang Wang, Gerard de Melo
| null | null | 2,020 |
aaai
|
Rethinking Generalization of Neural Models: A Named Entity Recognition Case Study
| null |
While neural network-based models have achieved impressive performance on a large body of NLP tasks, the generalization behavior of different models remains poorly understood: Does this excellent performance imply a perfect generalization model, or are there still some limitations? In this paper, we take the NER task as a testbed to analyze the generalization behavior of existing models from different perspectives and characterize the differences of their generalization abilities through the lens of our proposed measures, which guides us to better design models and training methods. Experiments with in-depth analyses diagnose the bottleneck of existing neural NER models in terms of breakdown performance analysis, annotation errors, dataset bias, and category relationships, which suggest directions for improvement. We have released the datasets: (ReCoNLL, PLONER) for the future research at our project page: http://pfliu.com/InterpretNER/.
|
Jinlan Fu, Pengfei Liu, Qi Zhang
| null | null | 2,020 |
aaai
|
Open Domain Event Text Generation
| null |
Text generation tasks aim at generating human-readable text from different kinds of data. Normally, the generated text only contains the information included in the data and its application is thus restricted to some limited scenarios. In this paper, we extend the task to an open domain event text generation scenario with an entity chain as its skeleton. Specifically, given an entity chain containing several related event entities, the model should retrieve from a trustworthy repository (e.g. Wikipedia) the detailed information of these entities and generate a description text based on the retrieved sentences. We build a new dataset called WikiEvent1 that provides 34K pairs of entity chain and its corresponding description sentences. To solve the problem, we propose a wiki augmented generator framework that contains an encoder, a retriever, and a decoder. The encoder encodes the entity chain into a hidden space while the decoder decodes from the hidden space and generates description text. The retriever retrieves relevant text from a trustworthy repository which provides more information for generation. To alleviate the overfitting problem, we propose a novel random drop component that randomly deletes words from the retrieved sentences making our model more robust for handling long input sentences. We apply the proposed model on the WikiEvent dataset and compare it with a few baselines. The experimental results show that our carefully-designed architecture does help generate better event text, and extensive analysis further uncovers the characteristics of the proposed task.
|
Zihao Fu, Lidong Bing, Wai Lam
| null | null | 2,020 |
aaai
|
Learning to Select Bi-Aspect Information for Document-Scale Text Content Manipulation
| null |
In this paper, we focus on a new practical task, document-scale text content manipulation, which is the opposite of text style transfer and aims to preserve text styles while altering the content. In detail, the input is a set of structured records and a reference text for describing another recordset. The output is a summary that accurately describes the partial content in the source recordset with the same writing style of the reference. The task is unsupervised due to lack of parallel data, and is challenging to select suitable records and style words from bi-aspect inputs respectively and generate a high-fidelity long document. To tackle those problems, we first build a dataset based on a basketball game report corpus as our testbed, and present an unsupervised neural model with interactive attention mechanism, which is used for learning the semantic relationship between records and reference texts to achieve better content transfer and better style preservation. In addition, we also explore the effectiveness of the back-translation in our task for constructing some pseudo-training pairs. Empirical results show superiority of our approaches over competitive methods, and the models also yield a new state-of-the-art result on a sentence-level dataset. 1
|
Xiaocheng Feng, Yawei Sun, Bing Qin, Heng Gong, Yibo Sun, Wei Bi, XiaoJiang Liu, Ting Liu
| null | null | 2,020 |
aaai
|
Translucent Answer Predictions in Multi-Hop Reading Comprehension
| null |
Research on the task of Reading Comprehension style Question Answering (RCQA) has gained momentum in recent years due to the emergence of human annotated datasets and associated leaderboards, for example CoQA, HotpotQA, SQuAD, TriviaQA, etc. While state-of-the-art has advanced considerably, there is still ample opportunity to advance it further on some important variants of the RCQA task. In this paper, we propose a novel deep neural architecture, called TAP (Translucent Answer Prediction), to identify answers and evidence (in the form of supporting facts) in an RCQA task requiring multi-hop reasoning. TAP comprises two loosely coupled networks – Local and Global Interaction eXtractor (LoGIX) and Answer Predictor (AP). LoGIX predicts supporting facts, whereas AP consumes these predicted supporting facts to predict the answer span. The novel design of LoGIX is inspired by two key design desiderata – local context and global interaction– that we identified by analyzing examples of multi-hop RCQA task. The loose coupling between LoGIX and the AP reveals the set of sentences used by the AP in predicting an answer. Therefore, answer predictions of TAP can be interpreted in a translucent manner. TAP offers state-of-the-art performance on the HotpotQA (Yang et al. 2018) dataset – an apt dataset for multi-hop RCQA task – as it occupies Rank-1 on its leaderboard (https://hotpotqa.github.io/) at the time of submission.
|
G P Shrivatsa Bhargav, Michael Glass, Dinesh Garg, Shirish Shevade, Saswati Dana, Dinesh Khandelwal, L Venkata Subramaniam, Alfio Gliozzo
| null | null | 2,020 |
aaai
|
Asymmetrical Hierarchical Networks with Attentive Interactions for Interpretable Review-Based Recommendation
| null |
Recently, recommender systems have been able to emit substantially improved recommendations by leveraging user-provided reviews. Existing methods typically merge all reviews of a given user (item) into a long document, and then process user and item documents in the same manner. In practice, however, these two sets of reviews are notably different: users' reviews reflect a variety of items that they have bought and are hence very heterogeneous in their topics, while an item's reviews pertain only to that single item and are thus topically homogeneous. In this work, we develop a novel neural network model that properly accounts for this important difference by means of asymmetric attentive modules. The user module learns to attend to only those signals that are relevant with respect to the target item, whereas the item module learns to extract the most salient contents with regard to properties of the item. Our multi-hierarchical paradigm accounts for the fact that neither are all reviews equally useful, nor are all sentences within each review equally pertinent. Extensive experimental results on a variety of real datasets demonstrate the effectiveness of our method.
|
Xin Dong, Jingchao Ni, Wei Cheng, Zhengzhang Chen, Bo Zong, Dongjin Song, Yanchi Liu, Haifeng Chen, Gerard de Melo
| null | null | 2,020 |
aaai
|
TANDA: Transfer and Adapt Pre-Trained Transformer Models for Answer Sentence Selection
| null |
We propose TandA, an effective technique for fine-tuning pre-trained Transformer models for natural language tasks. Specifically, we first transfer a pre-trained model into a model for a general task by fine-tuning it with a large and high-quality dataset. We then perform a second fine-tuning step to adapt the transferred model to the target domain. We demonstrate the benefits of our approach for answer sentence selection, which is a well-known inference task in Question Answering. We built a large scale dataset to enable the transfer step, exploiting the Natural Questions dataset. Our approach establishes the state of the art on two well-known benchmarks, WikiQA and TREC-QA, achieving the impressive MAP scores of 92% and 94.3%, respectively, which largely outperform the the highest scores of 83.4% and 87.5% of previous work. We empirically show that TandA generates more stable and robust models reducing the effort required for selecting optimal hyper-parameters. Additionally, we show that the transfer step of TandA makes the adaptation step more robust to noise. This enables a more effective use of noisy datasets for fine-tuning. Finally, we also confirm the positive impact of TandA in an industrial setting, using domain specific datasets subject to different types of noise.
|
Siddhant Garg, Thuy Vu, Alessandro Moschitti
| null | null | 2,020 |
aaai
|
Document Summarization with VHTM: Variational Hierarchical Topic-Aware Mechanism
| null |
Automatic text summarization focuses on distilling summary information from texts. This research field has been considerably explored over the past decades because of its significant role in many natural language processing tasks; however, two challenging issues block its further development: (1) how to yield a summarization model embedding topic inference rather than extending with a pre-trained one and (2) how to merge the latent topics into diverse granularity levels. In this study, we propose a variational hierarchical model to holistically address both issues, dubbed VHTM. Different from the previous work assisted by a pre-trained single-grained topic model, VHTM is the first attempt to jointly accomplish summarization with topic inference via variational encoder-decoder and merge topics into multi-grained levels through topic embedding and attention. Comprehensive experiments validate the superior performance of VHTM compared with the baselines, accompanying with semantically consistent topics.
|
Xiyan Fu, Jun Wang, Jinghan Zhang, Jinmao Wei, Zhenglu Yang
| null | null | 2,020 |
aaai
|
Latent Emotion Memory for Multi-Label Emotion Classification
| null |
Identifying multiple emotions in a sentence is an important research topic. Existing methods usually model the problem as multi-label classification task. However, previous methods have two issues, limiting the performance of the task. First, these models do not consider prior emotion distribution in a sentence. Second, they fail to effectively capture the context information closely related to the corresponding emotion. In this paper, we propose a Latent Emotion Memory network (LEM) for multi-label emotion classification. The proposed model can learn the latent emotion distribution without external knowledge, and can effectively leverage it into the classification network. Experimental results on two benchmark datasets show that the proposed model outperforms strong baselines, achieving the state-of-the-art performance.
|
Hao Fei, Yue Zhang, Yafeng Ren, Donghong Ji
| null | null | 2,020 |
aaai
|
Detecting Asks in Social Engineering Attacks: Impact of Linguistic and Structural Knowledge
| null |
Social engineers attempt to manipulate users into undertaking actions such as downloading malware by clicking links or providing access to money or sensitive information. Natural language processing, computational sociolinguistics, and media-specific structural clues provide a means for detecting both the ask (e.g., buy gift card) and the risk/reward implied by the ask, which we call framing (e.g., lose your job, get a raise). We apply linguistic resources such as Lexical Conceptual Structure to tackle ask detection and also leverage structural clues such as links and their proximity to identified asks to improve confidence in our results. Our experiments indicate that the performance of ask detection, framing detection, and identification of the top ask is improved by linguistically motivated classes coupled with structural clues such as links. Our approach is implemented in a system that informs users about social engineering risk situations.
|
Bonnie Dorr, Archna Bhatia, Adam Dalton, Brodie Mather, Bryanna Hebenstreit, Sashank Santhanam, Zhuo Cheng, Samira Shaikh, Alan Zemel, Tomek Strzalkowski
| null | null | 2,020 |
aaai
|
A Large-Scale Dataset for Argument Quality Ranking: Construction and Analysis
| null |
Identifying the quality of free-text arguments has become an important task in the rapidly expanding field of computational argumentation. In this work, we explore the challenging task of argument quality ranking. To this end, we created a corpus of 30,497 arguments carefully annotated for point-wise quality, released as part of this work. To the best of our knowledge, this is the largest dataset annotated for point-wise argument quality, larger by a factor of five than previously released datasets. Moreover, we address the core issue of inducing a labeled score from crowd annotations by performing a comprehensive evaluation of different approaches to this problem. In addition, we analyze the quality dimensions that characterize this dataset. Finally, we present a neural method for argument quality ranking, which outperforms several baselines on our own dataset, as well as previous methods published for another dataset.
|
Shai Gretz, Roni Friedman, Edo Cohen-Karlik, Assaf Toledo, Dan Lahav, Ranit Aharonov, Noam Slonim
| null | null | 2,020 |
aaai
|
Two-Level Transformer and Auxiliary Coherence Modeling for Improved Text Segmentation
| null |
Breaking down the structure of long texts into semantically coherent segments makes the texts more readable and supports downstream applications like summarization and retrieval. Starting from an apparent link between text coherence and segmentation, we introduce a novel supervised model for text segmentation with simple but explicit coherence modeling. Our model – a neural architecture consisting of two hierarchically connected Transformer networks – is a multi-task learning model that couples the sentence-level segmentation objective with the coherence objective that differentiates correct sequences of sentences from corrupt ones. The proposed model, dubbed Coherence-Aware Text Segmentation (CATS), yields state-of-the-art segmentation performance on a collection of benchmark datasets. Furthermore, by coupling CATS with cross-lingual word embeddings, we demonstrate its effectiveness in zero-shot language transfer: it can successfully segment texts in languages unseen in training.
|
Goran Glavaš, Swapna Somasundaran
| null | null | 2,020 |
aaai
|
Likelihood Ratios and Generative Classifiers for Unsupervised Out-of-Domain Detection in Task Oriented Dialog
| null |
The task of identifying out-of-domain (OOD) input examples directly at test-time has seen renewed interest recently due to increased real world deployment of models. In this work, we focus on OOD detection for natural language sentence inputs to task-based dialog systems. Our findings are three-fold:First, we curate and release ROSTD (Real Out-of-Domain Sentences From Task-oriented Dialog) - a dataset of 4K OOD examples for the publicly available dataset from (Schuster et al. 2019). In contrast to existing settings which synthesize OOD examples by holding out a subset of classes, our examples were authored by annotators with apriori instructions to be out-of-domain with respect to the sentences in an existing dataset.Second, we explore likelihood ratio based approaches as an alternative to currently prevalent paradigms. Specifically, we reformulate and apply these approaches to natural language inputs. We find that they match or outperform the latter on all datasets, with larger improvements on non-artificial OOD benchmarks such as our dataset. Our ablations validate that specifically using likelihood ratios rather than plain likelihood is necessary to discriminate well between OOD and in-domain data.Third, we propose learning a generative classifier and computing a marginal likelihood (ratio) for OOD detection. This allows us to use a principled likelihood while at the same time exploiting training-time labels. We find that this approach outperforms both simple likelihood (ratio) based and other prior approaches. We are hitherto the first to investigate the use of generative classifiers for OOD detection at test-time.
|
Varun Gangal, Abhinav Arora, Arash Einolghozati, Sonal Gupta
| null | null | 2,020 |
aaai
|
Predictive Engagement: An Efficient Metric for Automatic Evaluation of Open-Domain Dialogue Systems
| null |
User engagement is a critical metric for evaluating the quality of open-domain dialogue systems. Prior work has focused on conversation-level engagement by using heuristically constructed features such as the number of turns and the total time of the conversation. In this paper, we investigate the possibility and efficacy of estimating utterance-level engagement and define a novel metric, predictive engagement, for automatic evaluation of open-domain dialogue systems. Our experiments demonstrate that (1) human annotators have high agreement on assessing utterance-level engagement scores; (2) conversation-level engagement scores can be predicted from properly aggregated utterance-level engagement scores. Furthermore, we show that the utterance-level engagement scores can be learned from data. These scores can be incorporated into automatic evaluation metrics for open-domain dialogue systems to improve the correlation with human judgements. This suggests that predictive engagement can be used as a real-time feedback for training better dialogue models.
|
Sarik Ghazarian, Ralph Weischedel, Aram Galstyan, Nanyun Peng
| null | null | 2,020 |
aaai
|
Posterior-GAN: Towards Informative and Coherent Response Generation with Posterior Generative Adversarial Network
| null |
Neural conversational models learn to generate responses by taking into account the dialog history. These models are typically optimized over the query-response pairs with a maximum likelihood estimation objective. However, the query-response tuples are naturally loosely coupled, and there exist multiple responses that can respond to a given query, which leads the conversational model learning burdensome. Besides, the general dull response problem is even worsened when the model is confronted with meaningless response training instances. Intuitively, a high-quality response not only responds to the given query but also links up to the future conversations, in this paper, we leverage the query-response-future turn triples to induce the generated responses that consider both the given context and the future conversations. To facilitate the modeling of these triples, we further propose a novel encoder-decoder based generative adversarial learning framework, Posterior Generative Adversarial Network (Posterior-GAN), which consists of a forward and a backward generative discriminator to cooperatively encourage the generated response to be informative and coherent by two complementary assessment perspectives. Experimental results demonstrate that our method effectively boosts the informativeness and coherence of the generated response on both automatic and human evaluation, which verifies the advantages of considering two assessment perspectives.
|
Shaoxiong Feng, Hongshen Chen, Kan Li, Dawei Yin
| null | null | 2,020 |
aaai
|
DMRM: A Dual-Channel Multi-Hop Reasoning Model for Visual Dialog
| null |
Visual Dialog is a vision-language task that requires an AI agent to engage in a conversation with humans grounded in an image. It remains a challenging task since it requires the agent to fully understand a given question before making an appropriate response not only from the textual dialog history, but also from the visually-grounded information. While previous models typically leverage single-hop reasoning or single-channel reasoning to deal with this complex multimodal reasoning task, which is intuitively insufficient. In this paper, we thus propose a novel and more powerful Dual-channel Multi-hop Reasoning Model for Visual Dialog, named DMRM. DMRM synchronously captures information from the dialog history and the image to enrich the semantic representation of the question by exploiting dual-channel reasoning. Specifically, DMRM maintains a dual channel to obtain the question- and history-aware image features and the question- and image-aware dialog history features by a mulit-hop reasoning process in each channel. Additionally, we also design an effective multimodal attention to further enhance the decoder to generate more accurate responses. Experimental results on the VisDial v0.9 and v1.0 datasets demonstrate that the proposed model is effective and outperforms compared models by a significant margin.
|
Feilong Chen, Fandong Meng, Jiaming Xu, Peng Li, Bo Xu, Jie Zhou
| null | null | 2,020 |
aaai
|
Schema-Guided Multi-Domain Dialogue State Tracking with Graph Attention Neural Networks
| null |
Dialogue state tracking (DST) aims at estimating the current dialogue state given all the preceding conversation. For multi-domain DST, the data sparsity problem is also a major obstacle due to the increased number of state candidates. Existing approaches generally predict the value for each slot independently and do not consider slot relations, which may aggravate the data sparsity problem. In this paper, we propose a Schema-guided multi-domain dialogue State Tracker with graph attention networks (SST) that predicts dialogue states from dialogue utterances and schema graphs which contain slot relations in edges. We also introduce a graph attention matching network to fuse information from utterances and graphs, and a recurrent graph attention network to control state updating. Experiment results show that our approach obtains new state-of-the-art performance on both MultiWOZ 2.0 and MultiWOZ 2.1 benchmarks.
|
Lu Chen, Boer Lv, Chi Wang, Su Zhu, Bowen Tan, Kai Yu
| null | null | 2,020 |
aaai
|
Improving Entity Linking by Modeling Latent Entity Type Information
| null |
Existing state of the art neural entity linking models employ attention-based bag-of-words context model and pre-trained entity embeddings bootstrapped from word embeddings to assess topic level context compatibility. However, the latent entity type information in the immediate context of the mention is neglected, which causes the models often link mentions to incorrect entities with incorrect type. To tackle this problem, we propose to inject latent entity type information into the entity embeddings based on pre-trained BERT. In addition, we integrate a BERT-based entity similarity score into the local context model of a state-of-the-art model to better capture latent entity type information. Our model significantly outperforms the state-of-the-art entity linking models on standard benchmark (AIDA-CoNLL). Detailed experiment analysis demonstrates that our model corrects most of the type errors produced by the direct baseline.
|
Shuang Chen, Jinpeng Wang, Feng Jiang, Chin-Yew Lin
| null | null | 2,020 |
aaai
|
Two Birds with One Stone: Investigating Invertible Neural Networks for Inverse Problems in Morphology
| null |
Most problems in natural language processing can be approximated as inverse problems such as analysis and generation at variety of levels from morphological (e.g., cat+Plural↔cats) to semantic (e.g., (call + 1 2)↔“Calculate one plus two.”). Although the tasks in both directions are closely related, general approach in the field has been to design separate models specific for each task. However, having one shared model for both tasks, would help the researchers exploit the common knowledge among these problems with reduced time and memory requirements. We investigate a specific class of neural networks, called Invertible Neural Networks (INNs) (Ardizzone et al. 2019) that enable simultaneous optimization in both directions, hence allow addressing of inverse problems via a single model. In this study, we investigate INNs on morphological problems casted as inverse problems. We apply INNs to various morphological tasks with varying ambiguity and show that they provide competitive performance in both directions. We show that they are able to recover the morphological input parameters, i.e., predicting the lemma (e.g., cat) or the morphological tags (e.g., Plural) when run in the reverse direction, without any significant performance drop in the forward direction, i.e., predicting the surface form (e.g., cats).
|
Gözde Gül Şahin, Iryna Gurevych
| null | null | 2,020 |
aaai
|
Working Memory-Driven Neural Networks with a Novel Knowledge Enhancement Paradigm for Implicit Discourse Relation Recognition
| null |
Recognizing implicit discourse relation is a challenging task in discourse analysis, which aims to understand and infer the latent relations between two discourse arguments, such as temporal, comparison. Most of the present models largely focus on learning-based methods that utilize only intra-sentence textual information to identify discourse relations, ignoring the wider contexts beyond the discourse. Moreover, people comprehend the meanings and the relations of discourses, heavily relying on their interconnected working memories (e.g., instant memory, long-term memory). Inspired by this, we propose a Knowledge-Enhanced Attentive Neural Network (KANN) framework to address these issues. Specifically, it establishes a mutual attention matrix to capture the reciprocal information between two arguments, as instant memory. While implicitly stated knowledge in the arguments is retrieved from external knowledge source and encoded as inter-words semantic connection embeddings to further construct knowledge matrix, as long-term memory. We devise a novel paradigm with two ways by the collaboration of the memories to enrich the argument representation: 1) integrating the knowledge matrix into the mutual attention matrix, which implicitly maps knowledge into the process of capturing asymmetric interactions between two discourse arguments; 2) directly concatenating the argument representations and the semantic connection embeddings, which explicitly supplements knowledge to help discourse understanding. The experimental results on the PDTB also show that our KANN model is effective.
|
Fengyu Guo, Ruifang He, Jianwu Dang, Jian Wang
| null | null | 2,020 |
aaai
|
TemPEST: Soft Template-Based Personalized EDM Subject Generation through Collaborative Summarization
| null |
We address personalized Electronic Direct Mail (EDM) subject generation, which generates an attractive subject line for a product description according to user's preference on different contents or writing styles. Generating personalized EDM subjects has a few notable differences from generating text summaries. The subject has to be not only faithful to the description itself but also attractive to increase the click-through rate. Moreover, different users may have different preferences over the styles of topics. We propose a novel personalized EDM subject generation model named Soft Template-based Personalized EDM Subject Generator (TemPEST) to consider the aforementioned users' characteristics when generating subjects, which contains a soft template-based selective encoder network, a user rating encoder network, a summary decoder network and a rating decoder. Experimental results indicate that TemPEST is able to generate personalized topics and also effectively perform recommending rating reconstruction.
|
Yu-Hsiu Chen, Pin-Yu Chen, Hong-Han Shuai, Wen-Chih Peng
| null | null | 2,020 |
aaai
|
Sequence Generation with Optimal-Transport-Enhanced Reinforcement Learning
| null |
Reinforcement learning (RL) has been widely used to aid training in language generation. This is achieved by enhancing standard maximum likelihood objectives with user-specified reward functions that encourage global semantic consistency. We propose a principled approach to address the difficulties associated with RL-based solutions, namely, high-variance gradients, uninformative rewards and brittle training. By leveraging the optimal transport distance, we introduce a regularizer that significantly alleviates the above issues. Our formulation emphasizes the preservation of semantic features, enabling end-to-end training instead of ad-hoc fine-tuning, and when combined with RL, it controls the exploration space for more efficient model updates. To validate the effectiveness of the proposed solution, we perform a comprehensive evaluation covering a wide variety of NLP tasks: machine translation, abstractive text summarization and image caption, with consistent improvements over competing solutions.
|
Liqun Chen, Ke Bai, Chenyang Tao, Yizhe Zhang, Guoyin Wang, Wenlin Wang, Ricardo Henao, Lawrence Carin
| null | null | 2,020 |
aaai
|
Learning to Map Frequent Phrases to Sub-Structures of Meaning Representation for Neural Semantic Parsing
| null |
Neural semantic parsers usually generate meaning representation tokens from natural language tokens via an encoder-decoder model. However, there is often a vocabulary-mismatch problem between natural language utterances and logical forms. That is, one word maps to several atomic logical tokens, which need to be handled as a whole, rather than individual logical tokens at multiple steps. In this paper, we propose that the vocabulary-mismatch problem can be effectively resolved by leveraging appropriate logical tokens. Specifically, we exploit macro actions, which are of the same granularity of words/phrases, and allow the model to learn mappings from frequent phrases to corresponding sub-structures of meaning representation. Furthermore, macro actions are compact, and therefore utilizing them can significantly reduce the search space, which brings a great benefit to weakly supervised semantic parsing. Experiments show that our method leads to substantial performance improvement on three benchmarks, in both supervised and weakly supervised settings.
|
Bo Chen, Xianpei Han, Ben He, Le Sun
| null | null | 2,020 |
aaai
|
Discriminative Sentence Modeling for Story Ending Prediction
| null |
Story Ending Prediction is a task that needs to select an appropriate ending for the given story, which requires the machine to understand the story and sometimes needs commonsense knowledge. To tackle this task, we propose a new neural network called Diff-Net for better modeling the differences of each ending in this task. The proposed model could discriminate two endings in three semantic levels: contextual representation, story-aware representation, and discriminative representation. Experimental results on the Story Cloze Test dataset show that the proposed model siginificantly outperforms various systems by a large margin, and detailed ablation studies are given for better understanding our model. We also carefully examine the traditional and BERT-based models on both SCT v1.0 and v1.5 with interesting findings that may potentially help future studies.
|
Yiming Cui, Wanxiang Che, Wei-Nan Zhang, Ting Liu, Shijin Wang, Guoping Hu
| null | null | 2,020 |
aaai
|
An Empirical Study of Content Understanding in Conversational Question Answering
| null |
With a lot of work about context-free question answering systems, there is an emerging trend of conversational question answering models in the natural language processing field. Thanks to the recently collected datasets, including QuAC and CoQA, there has been more work on conversational question answering, and recent work has achieved competitive performance on both datasets. However, to best of our knowledge, two important questions for conversational comprehension research have not been well studied: 1) How well can the benchmark dataset reflect models' content understanding? 2) Do the models well utilize the conversation content when answering questions? To investigate these questions, we design different training settings, testing settings, as well as an attack to verify the models' capability of content understanding on QuAC and CoQA. The experimental results indicate some potential hazards in the benchmark datasets, QuAC and CoQA, for conversational comprehension research. Our analysis also sheds light on both what models may learn and how datasets may bias the models. With deep investigation of the task, it is believed that this work can benefit the future progress of conversation comprehension. The source code is available at https://github.com/MiuLab/CQA-Study.
|
Ting-Rui Chiang, Hao-Tong Ye, Yun-Nung Chen
| null | null | 2,020 |
aaai
|
How to Ask Better Questions? A Large-Scale Multi-Domain Dataset for Rewriting Ill-Formed Questions
| null |
We present a large-scale dataset for the task of rewriting an ill-formed natural language question to a well-formed one. Our multi-domain question rewriting (MQR) dataset is constructed from human contributed Stack Exchange question edit histories. The dataset contains 427,719 question pairs which come from 303 domains. We provide human annotations for a subset of the dataset as a quality estimate. When moving from ill-formed to well-formed questions, the question quality improves by an average of 45 points across three aspects. We train sequence-to-sequence neural models on the constructed dataset and obtain an improvement of 13.2% in BLEU-4 over baseline methods built from other data resources. We release the MQR dataset to encourage research on the problem of question rewriting.1
|
Zewei Chu, Mingda Chen, Jing Chen, Miaosen Wang, Kevin Gimpel, Manaal Faruqui, Xiance Si
| null | null | 2,020 |
aaai
|
Guiding Attention in Sequence-to-Sequence Models for Dialogue Act Prediction
| null |
The task of predicting dialog acts (DA) based on conversational dialog is a key component in the development of conversational agents. Accurately predicting DAs requires a precise modeling of both the conversation and the global tag dependencies. We leverage seq2seq approaches widely adopted in Neural Machine Translation (NMT) to improve the modelling of tag sequentiality. Seq2seq models are known to learn complex global dependencies while currently proposed approaches using linear conditional random fields (CRF) only model local tag dependencies. In this work, we introduce a seq2seq model tailored for DA classification using: a hierarchical encoder, a novel guided attention mechanism and beam search applied to both training and inference. Compared to the state of the art our model does not require handcrafted features and is trained end-to-end. Furthermore, the proposed approach achieves an unmatched accuracy score of 85% on SwDA, and state-of-the-art accuracy score of 91.6% on MRDA.
|
Pierre Colombo, Emile Chapuis, Matteo Manica, Emmanuel Vignon, Giovanna Varni, Chloe Clavel
| null | null | 2,020 |
aaai
|
Cross-Lingual Natural Language Generation via Pre-Training
| null |
In this work we focus on transferring supervision signals of natural language generation (NLG) tasks between multiple languages. We propose to pretrain the encoder and the decoder of a sequence-to-sequence model under both monolingual and cross-lingual settings. The pre-training objective encourages the model to represent different languages in the shared space, so that we can conduct zero-shot cross-lingual transfer. After the pre-training procedure, we use monolingual data to fine-tune the pre-trained model on downstream NLG tasks. Then the sequence-to-sequence model trained in a single language can be directly evaluated beyond that language (i.e., accepting multi-lingual input and producing multi-lingual output). Experimental results on question generation and abstractive summarization show that our model outperforms the machine-translation-based pipeline methods for zero-shot cross-lingual generation. Moreover, cross-lingual transfer improves NLG performance of low-resource languages by leveraging rich-resource language data. Our implementation and data are available at https://github.com/CZWin32768/xnlg.
|
Zewen Chi, Li Dong, Furu Wei, Wenhui Wang, Xian-Ling Mao, Heyan Huang
| null | null | 2,020 |
aaai
|
Multiple Positional Self-Attention Network for Text Classification
| null |
Self-attention mechanisms have recently caused many concerns on Natural Language Processing (NLP) tasks. Relative positional information is important to self-attention mechanisms. We propose Faraway Mask focusing on the (2m + 1)-gram words and Scaled-Distance Mask putting the logarithmic distance punishment to avoid and weaken the self-attention of distant words respectively. To exploit different masks, we present Positional Self-Attention Layer for generating different Masked-Self-Attentions and a following Position-Fusion Layer in which fused positional information multiplies the Masked-Self-Attentions for generating sentence embeddings. To evaluate our sentence embeddings approach Multiple Positional Self-Attention Network (MPSAN), we perform the comparison experiments on sentiment analysis, semantic relatedness and sentence classification tasks. The result shows that our MPSAN outperforms state-of-the-art methods on five datasets and the test accuracy is improved by 0.81%, 0.6% on SST, CR datasets, respectively. In addition, we reduce training parameters and improve the time efficiency of MPSAN by lowering the dimension number of self-attention and simplifying fusion mechanism.
|
Biyun Dai, Jinlong Li, Ruoyi Xu
| null | null | 2,020 |
aaai
|
Attending to Entities for Better Text Understanding
| null |
Recent progress in NLP witnessed the development of large-scale pre-trained language models (GPT, BERT, XLNet, etc.) based on Transformer (Vaswani et al. 2017), and in a range of end tasks, such models have achieved state-of-the-art results, approaching human performance. This clearly demonstrates the power of the stacked self-attention architecture when paired with a sufficient number of layers and a large amount of pre-training data. However, on tasks that require complex and long-distance reasoning where surface-level cues are not enough, there is still a large gap between the pre-trained models and human performance. Strubell et al. (2018) recently showed that it is possible to inject knowledge of syntactic structure into a model through supervised self-attention. We conjecture that a similar injection of semantic knowledge, in particular, coreference information, into an existing model would improve performance on such complex problems. On the LAMBADA (Paperno et al. 2016) task, we show that a model trained from scratch with coreference as auxiliary supervision for self-attention outperforms the largest GPT-2 model, setting the new state-of-the-art, while only containing a tiny fraction of parameters compared to GPT-2. We also conduct a thorough analysis of different variants of model architectures and supervision configurations, suggesting future directions on applying similar techniques to other problems.
|
Pengxiang Cheng, Katrin Erk
| null | null | 2,020 |
aaai
|
Understanding the Semantic Content of Sparse Word Embeddings Using a Commonsense Knowledge Base
| null |
Word embeddings have developed into a major NLP tool with broad applicability. Understanding the semantic content of word embeddings remains an important challenge for additional applications. One aspect of this issue is to explore the interpretability of word embeddings. Sparse word embeddings have been proposed as models with improved interpretability. Continuing this line of research, we investigate the extent to which human interpretable semantic concepts emerge along the bases of sparse word representations. In order to have a broad framework for evaluation, we consider three general approaches for constructing sparse word representations, which are then evaluated in multiple ways. We propose a novel methodology to evaluate the semantic content of word embeddings using a commonsense knowledge base, applied here to the sparse case. This methodology is illustrated by two techniques using the ConceptNet knowledge base. The first approach assigns a commonsense concept label to the individual dimensions of the embedding space. The second approach uses a metric, derived by spreading activation, to quantify the coherence of coordinates along the individual axes. We also provide results on the relationship between the two approaches. The results show, for example, that in the individual dimensions of sparse word embeddings, words having high coefficients are more semantically related in terms of path lengths in the knowledge base than the ones having zero coefficients.
|
Vanda Balogh, Gábor Berend, Dimitrios I. Diochnos, György Turán
| null | null | 2,020 |
aaai
|
Do Not Have Enough Data? Deep Learning to the Rescue!
| null |
Based on recent advances in natural language modeling and those in text generation capabilities, we propose a novel data augmentation method for text classification tasks. We use a powerful pre-trained neural network model to artificially synthesize new labeled data for supervised learning. We mainly focus on cases with scarce labeled data. Our method, referred to as language-model-based data augmentation (LAMBADA), involves fine-tuning a state-of-the-art language generator to a specific task through an initial training phase on the existing (usually small) labeled data. Using the fine-tuned model and given a class label, new sentences for the class are generated. Our process then filters these new sentences by using a classifier trained on the original data. In a series of experiments, we show that LAMBADA improves classifiers' performance on a variety of datasets. Moreover, LAMBADA significantly improves upon the state-of-the-art techniques for data augmentation, specifically those applicable to text classification tasks with little data.
|
Ateret Anaby-Tavor, Boaz Carmeli, Esther Goldbraich, Amir Kantor, George Kour, Segev Shlomov, Naama Tepper, Naama Zwerdling
| null | null | 2,020 |
aaai
|
Simultaneously Linking Entities and Extracting Relations from Biomedical Text without Mention-Level Supervision
| null |
Understanding the meaning of text often involves reasoning about entities and their relationships. This requires identifying textual mentions of entities, linking them to a canonical concept, and discerning their relationships. These tasks are nearly always viewed as separate components within a pipeline, each requiring a distinct model and training data. While relation extraction can often be trained with readily available weak or distant supervision, entity linkers typically require expensive mention-level supervision – which is not available in many domains. Instead, we propose a model which is trained to simultaneously produce entity linking and relation decisions while requiring no mention-level annotations. This approach avoids cascading errors that arise from pipelined methods and more accurately predicts entity relationships from text. We show that our model outperforms a state-of-the art entity linking and relation extraction pipeline on two biomedical datasets and can drastically improve the overall recall of the system.
|
Trapit Bansal, Pat Verga, Neha Choudhary, Andrew McCallum
| null | null | 2,020 |
aaai
|
Fine-Grained Named Entity Typing over Distantly Supervised Data Based on Refined Representations
| null |
Fine-Grained Named Entity Typing (FG-NET) is a key component in Natural Language Processing (NLP). It aims at classifying an entity mention into a wide range of entity types. Due to a large number of entity types, distant supervision is used to collect training data for this task, which noisily assigns type labels to entity mentions irrespective of the context. In order to alleviate the noisy labels, existing approaches on FG-NET analyze the entity mentions entirely independent of each other and assign type labels solely based on mention's sentence-specific context. This is inadequate for highly overlapping and/or noisy type labels as it hinders information passing across sentence boundaries. For this, we propose an edge-weighted attentive graph convolution network that refines the noisy mention representations by attending over corpus-level contextual clues prior to the end classification. Experimental evaluation shows that the proposed model outperforms the existing research by a relative score of upto 10.2% and 8.3% for macro-f1 and micro-f1 respectively.
|
Muhammad Asif Ali, Yifang Sun, Bing Li, Wei Wang
| null | null | 2,020 |
aaai
|
Joint Learning of Answer Selection and Answer Summary Generation in Community Question Answering
| null |
Community question answering (CQA) gains increasing popularity in both academy and industry recently. However, the redundancy and lengthiness issues of crowdsourced answers limit the performance of answer selection and lead to reading difficulties and misunderstandings for community users. To solve these problems, we tackle the tasks of answer selection and answer summary generation in CQA with a novel joint learning model. Specifically, we design a question-driven pointer-generator network, which exploits the correlation information between question-answer pairs to aid in attending the essential information when generating answer summaries. Meanwhile, we leverage the answer summaries to alleviate noise in original lengthy answers when ranking the relevancy degrees of question-answer pairs. In addition, we construct a new large-scale CQA corpus, WikiHowQA, which contains long answers for answer selection as well as reference summaries for answer summarization. The experimental results show that the joint learning method can effectively address the answer redundancy issue in CQA and achieves state-of-the-art results on both answer selection and text summarization tasks. Furthermore, the proposed model is shown to be of great transferring ability and applicability for resource-poor CQA tasks, which lack of reference answer summaries.
|
Yang Deng, Wai Lam, Yuexiang Xie, Daoyuan Chen, Yaliang Li, Min Yang, Ying Shen
| null | null | 2,020 |
aaai
|
End-to-End Argumentation Knowledge Graph Construction
| null |
This paper studies the end-to-end construction of an argumentation knowledge graph that is intended to support argument synthesis, argumentative question answering, or fake news detection, among others. The study is motivated by the proven effectiveness of knowledge graphs for interpretable and controllable text generation and exploratory search. Original in our work is that we propose a model of the knowledge encapsulated in arguments. Based on this model, we build a new corpus that comprises about 16k manual annotations of 4740 claims with instances of the model's elements, and we develop an end-to-end framework that automatically identifies all modeled types of instances. The results of experiments show the potential of the framework for building a web-based argumentation graph that is of high quality and large scale.
|
Khalid Al-Khatib, Yufang Hou, Henning Wachsmuth, Charles Jochim, Francesca Bonin, Benno Stein
| null | null | 2,020 |
aaai
|
Modelling Sentence Pairs via Reinforcement Learning: An Actor-Critic Approach to Learn the Irrelevant Words
| null |
Learning sentence representation is a fundamental task in Natural Language Processing. Most of the existing sentence pair modelling architectures focus only on extracting and using the rich sentence pair features. The drawback of utilizing all of these features makes the learning process much harder. In this study, we propose a reinforcement learning (RL) method to learn a sentence pair representation when performing tasks like semantic similarity, paraphrase identification, and question-answer pair modelling. We formulate this learning problem as a sequential decision making task where the decision made in the current state will have a strong impact on the following decisions. We address this decision making with a policy gradient RL method which chooses the irrelevant words to delete by looking at the sub-optimal representation of the sentences being compared. With this policy, extensive experiments show that our model achieves on par performance when learning task-specific representations of sentence pairs without needing any further knowledge like parse trees. We suggest that the simplicity of each task inference provided by our RL model makes it easier to explain.
|
Mahtab Ahmed, Robert E. Mercer
| null | null | 2,020 |
aaai
|
LeDeepChef Deep Reinforcement Learning Agent for Families of Text-Based Games
| null |
While Reinforcement Learning (RL) approaches lead to significant achievements in a variety of areas in recent history, natural language tasks remained mostly unaffected, due to the compositional and combinatorial nature that makes them notoriously hard to optimize. With the emerging field of Text-Based Games (TBGs), researchers try to bridge this gap. Inspired by the success of RL algorithms on Atari games, the idea is to develop new methods in a restricted game world and then gradually move to more complex environments. Previous work in the area of TBGs has mainly focused on solving individual games. We, however, consider the task of designing an agent that not just succeeds in a single game, but performs well across a whole family of games, sharing the same theme. In this work, we present our deep RL agent—LeDeepChef—that shows generalization capabilities to never-before-seen games of the same family with different environments and task descriptions. The agent participated in Microsoft Research's First TextWorld Problems: A Language and Reinforcement Learning Challenge and outperformed all but one competitor on the final test set. The games from the challenge all share the same theme, namely cooking in a modern house environment, but differ significantly in the arrangement of the rooms, the presented objects, and the specific goal (recipe to cook). To build an agent that achieves high scores across a whole family of games, we use an actor-critic framework and prune the action-space by using ideas from hierarchical reinforcement learning and a specialized module trained on a recipe database.
|
Leonard Adolphs, Thomas Hofmann
| null | null | 2,020 |
aaai
|
An Iterative Polishing Framework Based on Quality Aware Masked Language Model for Chinese Poetry Generation
| null |
Owing to its unique literal and aesthetical characteristics, automatic generation of Chinese poetry is still challenging in Artificial Intelligence, which can hardly be straightforwardly realized by end-to-end methods. In this paper, we propose a novel iterative polishing framework for highly qualified Chinese poetry generation. In the first stage, an encoder-decoder structure is utilized to generate a poem draft. Afterwards, our proposed Quality-Aware Masked Language Model (QA-MLM) is employed to polish the draft towards higher quality in terms of linguistics and literalness. Based on a multi-task learning scheme, QA-MLM is able to determine whether polishing is needed based on the poem draft. Furthermore, QA-MLM is able to localize improper characters of the poem draft and substitute with newly predicted ones accordingly. Benefited from the masked language model structure, QA-MLM incorporates global context information into the polishing process, which can obtain more appropriate polishing results than the unidirectional sequential decoding. Moreover, the iterative polishing process will be terminated automatically when QA-MLM regards the processed poem as a qualified one. Both human and automatic evaluation have been conducted, and the results demonstrate that our approach is effective to improve the performance of encoder-decoder structure.
|
Liming Deng, Jie Wang, Hangming Liang, Hui Chen, Zhiqiang Xie, Bojin Zhuang, Shaojun Wang, Jing Xiao
| null | null | 2,020 |
aaai
|
Knowledge Distillation from Internal Representations
| null |
Knowledge distillation is typically conducted by training a small model (the student) to mimic a large and cumbersome model (the teacher). The idea is to compress the knowledge from the teacher by using its output probabilities as soft-labels to optimize the student. However, when the teacher is considerably large, there is no guarantee that the internal knowledge of the teacher will be transferred into the student; even if the student closely matches the soft-labels, its internal representations may be considerably different. This internal mismatch can undermine the generalization capabilities originally intended to be transferred from the teacher to the student. In this paper, we propose to distill the internal representations of a large model such as BERT into a simplified version of it. We formulate two ways to distill such representations and various algorithms to conduct the distillation. We experiment with datasets from the GLUE benchmark and consistently show that adding knowledge distillation from internal representations is a more powerful method than only using soft-label distillation.
|
Gustavo Aguilar, Yuan Ling, Yu Zhang, Benjamin Yao, Xing Fan, Chenlei Guo
| null | null | 2,020 |
aaai
|
Adversarial Training Based Multi-Source Unsupervised Domain Adaptation for Sentiment Analysis
| null |
Multi-source unsupervised domain adaptation (MS-UDA) for sentiment analysis (SA) aims to leverage useful information in multiple source domains to help do SA in an unlabeled target domain that has no supervised information. Existing algorithms of MS-UDA either only exploit the shared features, i.e., the domain-invariant information, or based on some weak assumption in NLP, e.g., smoothness assumption. To avoid these problems, we propose two transfer learning frameworks based on the multi-source domain adaptation methodology for SA by combining the source hypotheses to derive a good target hypothesis. The key feature of the first framework is a novel Weighting Scheme based Unsupervised Domain Adaptation framework ((WS-UDA), which combine the source classifiers to acquire pseudo labels for target instances directly. While the second framework is a Two-Stage Training based Unsupervised Domain Adaptation framework (2ST-UDA), which further exploits these pseudo labels to train a target private extractor. Importantly, the weights assigned to each source classifier are based on the relations between target instances and source domains, which measured by a discriminator through the adversarial training. Furthermore, through the same discriminator, we also fulfill the separation of shared features and private features.Experimental results on two SA datasets demonstrate the promising performance of our frameworks, which outperforms unsupervised state-of-the-art competitors.
|
Yong Dai, Jian Liu, Xiancong Ren, Zenglin Xu
| null | null | 2,020 |
aaai
|
Modelling Semantic Categories Using Conceptual Neighborhood
| null |
While many methods for learning vector space embeddings have been proposed in the field of Natural Language Processing, these methods typically do not distinguish between categories and individuals. Intuitively, if individuals are represented as vectors, we can think of categories as (soft) regions in the embedding space. Unfortunately, meaningful regions can be difficult to estimate, especially since we often have few examples of individuals that belong to a given category. To address this issue, we rely on the fact that different categories are often highly interdependent. In particular, categories often have conceptual neighbors, which are disjoint from but closely related to the given category (e.g. fruit and vegetable). Our hypothesis is that more accurate category representations can be learned by relying on the assumption that the regions representing such conceptual neighbors should be adjacent in the embedding space. We propose a simple method for identifying conceptual neighbors and then show that incorporating these conceptual neighbors indeed leads to more accurate region based representations.
|
Zied Bouraoui, Jose Camacho-Collados, Luis Espinosa-Anke, Steven Schockaert
| null | null | 2,020 |
aaai
|
Learning from Easy to Complex: Adaptive Multi-Curricula Learning for Neural Dialogue Generation
| null |
Current state-of-the-art neural dialogue systems are mainly data-driven and are trained on human-generated responses. However, due to the subjectivity and open-ended nature of human conversations, the complexity of training dialogues varies greatly. The noise and uneven complexity of query-response pairs impede the learning efficiency and effects of the neural dialogue generation models. What is more, so far, there are no unified dialogue complexity measurements, and the dialogue complexity embodies multiple aspects of attributes—specificity, repetitiveness, relevance, etc. Inspired by human behaviors of learning to converse, where children learn from easy dialogues to complex ones and dynamically adjust their learning progress, in this paper, we first analyze five dialogue attributes to measure the dialogue complexity in multiple perspectives on three publicly available corpora. Then, we propose an adaptive multi-curricula learning framework to schedule a committee of the organized curricula. The framework is established upon the reinforcement learning paradigm, which automatically chooses different curricula at the evolving learning process according to the learning status of the neural dialogue generation model. Extensive experiments conducted on five state-of-the-art models demonstrate its learning efficiency and effectiveness with respect to 13 automatic evaluation metrics and human judgments.
|
Hengyi Cai, Hongshen Chen, Cheng Zhang, Yonghao Song, Xiaofang Zhao, Yangxi Li, Dongsheng Duan, Dawei Yin
| null | null | 2,020 |
aaai
|
Back to the Future – Temporal Adaptation of Text Representations
| null |
Language evolves over time in many ways relevant to natural language processing tasks. For example, recent occurrences of tokens 'BERT' and 'ELMO' in publications refer to neural network architectures rather than persons. This type of temporal signal is typically overlooked, but is important if one aims to deploy a machine learning model over an extended period of time. In particular, language evolution causes data drift between time-steps in sequential decision-making tasks. Examples of such tasks include prediction of paper acceptance for yearly conferences (regular intervals) or author stance prediction for rumours on Twitter (irregular intervals). Inspired by successes in computer vision, we tackle data drift by sequentially aligning learned representations. We evaluate on three challenging tasks varying in terms of time-scales, linguistic units, and domains. These tasks show our method outperforming several strong baselines, including using all available data. We argue that, due to its low computational expense, sequential alignment is a practical solution to dealing with language evolution.
|
Johannes Bjerva, Wouter Kouw, Isabelle Augenstein
| null | null | 2,020 |
aaai
|
Symbolic Top-k Planning
| null |
The objective of top-k planning is to determine a set of k different plans with lowest cost for a given planning task. In practice, such a set of best plans can be preferred to a single best plan generated by ordinary optimal planners, as it allows the user to choose between different alternatives and thus take into account preferences that may be difficult to model. In this paper we show that, in general, the decision problem version of top-k planning is PSPACE-complete, as is the decision problem version of ordinary classical planning. This does not hold for polynomially bounded plans for which the decision problem turns out to be PP-hard, while the ordinary case is NP-hard. We present a novel approach to top-k planning, called sym-k, which is based on symbolic search, and prove that sym-k is sound and complete. Our empirical analysis shows that sym-k exceeds the current state of the art for both small and large k.
|
David Speck, Robert Mattmüller, Bernhard Nebel
| null | null | 2,020 |
aaai
|
Inducing Relational Knowledge from BERT
| null |
One of the most remarkable properties of word embeddings is the fact that they capture certain types of semantic and syntactic relationships. Recently, pre-trained language models such as BERT have achieved groundbreaking results across a wide range of Natural Language Processing tasks. However, it is unclear to what extent such models capture relational knowledge beyond what is already captured by standard word embeddings. To explore this question, we propose a methodology for distilling relational knowledge from a pre-trained language model. Starting from a few seed instances of a given relation, we first use a large text corpus to find sentences that are likely to express this relation. We then use a subset of these extracted sentences as templates. Finally, we fine-tune a language model to predict whether a given word pair is likely to be an instance of some relation, when given an instantiated template for that relation as input.
|
Zied Bouraoui, Jose Camacho-Collados, Steven Schockaert
| null | null | 2,020 |
aaai
|
Generating Well-Formed Answers by Machine Reading with Stochastic Selector Networks
| null |
Question answering (QA) based on machine reading comprehension has been a recent surge in popularity, yet most work has focused on extractive methods. We instead address a more challenging QA problem of generating a well-formed answer by reading and summarizing the paragraph for a given question.For the generative QA task, we introduce a new neural architecture, LatentQA, in which a novel stochastic selector network composes a well-formed answer with words selected from the question, the paragraph and the global vocabulary, based on a sequence of discrete latent variables. Bayesian inference for the latent variables is performed to train the LatentQA model. The experiments on public datasets of natural answer generation confirm the effectiveness of LatentQA in generating high-quality well-formed answers.
|
Bin Bi, Chen Wu, Ming Yan, Wei Wang, Jiangnan Xia, Chenliang Li
| null | null | 2,020 |
aaai
|
Hyperbolic Interaction Model for Hierarchical Multi-Label Classification
| null |
Different from the traditional classification tasks which assume mutual exclusion of labels, hierarchical multi-label classification (HMLC) aims to assign multiple labels to every instance with the labels organized under hierarchical relations. Besides the labels, since linguistic ontologies are intrinsic hierarchies, the conceptual relations between words can also form hierarchical structures. Thus it can be a challenge to learn mappings from word hierarchies to label hierarchies. We propose to model the word and label hierarchies by embedding them jointly in the hyperbolic space. The main reason is that the tree-likeness of the hyperbolic space matches the complexity of symbolic data with hierarchical structures. A new Hyperbolic Interaction Model (HyperIM) is designed to learn the label-aware document representations and make predictions for HMLC. Extensive experiments are conducted on three benchmark datasets. The results have demonstrated that the new model can realistically capture the complex data structures and further improve the performance for HMLC comparing with the state-of-the-art methods. To facilitate future research, our code is publicly available.
|
Boli Chen, Xin Huang, Lin Xiao, Zixin Cai, Liping Jing
| null | null | 2,020 |
aaai
|
Graph Transformer for Graph-to-Sequence Learning
| null |
The dominant graph-to-sequence transduction models employ graph neural networks for graph representation learning, where the structural information is reflected by the receptive field of neurons. Unlike graph neural networks that restrict the information exchange between immediate neighborhood, we propose a new model, known as Graph Transformer, that uses explicit relation encoding and allows direct communication between two distant nodes. It provides a more efficient way for global graph structure modeling. Experiments on the applications of text generation from Abstract Meaning Representation (AMR) and syntax-based neural machine translation show the superiority of our proposed model. Specifically, our model achieves 27.4 BLEU on LDC2015E86 and 29.7 BLEU on LDC2017T10 for AMR-to-text generation, outperforming the state-of-the-art results by up to 2.2 points. On the syntax-based translation tasks, our model establishes new single-model state-of-the-art BLEU scores, 21.3 for English-to-German and 14.1 for English-to-Czech, improving over the existing best results, including ensembles, by over 1 BLEU.
|
Deng Cai, Wai Lam
| null | null | 2,020 |
aaai
|
Active Goal Recognition
| null |
The objective of goal recognition is to infer a goal that accounts for the observed behavior of an actor. In this work, we introduce and formalize the notion of active goal recognition in which we endow the observer with agency to sense, reason, and act in the world with a view to enhancing and possibly expediting goal recognition, and/or to intervening in goal achievement. To this end, we present an algorithm for active goal recognition and a landmark-based approach to the elimination of hypothesized goals which leverages automated planning. Experiments demonstrate the merits of providing agency to the observer, and the effectiveness of our approach in potentially enhancing the observational power of the observer, as well as expediting and in some cases making possible the recognition of the actor's goal.
|
Maayan Shvo, Sheila A. McIlraith
| null | null | 2,020 |
aaai
|
NeoNav: Improving the Generalization of Visual Navigation via Generating Next Expected Observations
| null |
We propose improving the cross-target and cross-scene generalization of visual navigation through learning an agent that is guided by conceiving the next observations it expects to see. This is achieved by learning a variational Bayesian model, called NeoNav, which generates the next expected observations (NEO) conditioned on the current observations of the agent and the target view. Our generative model is learned through optimizing a variational objective encompassing two key designs. First, the latent distribution is conditioned on current observations and the target view, leading to a model-based, target-driven navigation. Second, the latent space is modeled with a Mixture of Gaussians conditioned on the current observation and the next best action. Our use of mixture-of-posteriors prior effectively alleviates the issue of over-regularized latent space, thus significantly boosting the model generalization for new targets and in novel scenes. Moreover, the NEO generation models the forward dynamics of agent-environment interaction, which improves the quality of approximate inference and hence benefits data efficiency. We have conducted extensive evaluations on both real-world and synthetic benchmarks, and show that our model consistently outperforms the state-of-the-art models in terms of success rate, data efficiency, and generalization.
|
Qiaoyun Wu, Dinesh Manocha, Jun Wang, Kai Xu
| null | null | 2,020 |
aaai
|
Zero-Shot Text-to-SQL Learning with Auxiliary Task
| null |
Recent years have seen great success in the use of neural seq2seq models on the text-to-SQL task. However, little work has paid attention to how these models generalize to realistic unseen data, which naturally raises a question: does this impressive performance signify a perfect generalization model, or are there still some limitations?In this paper, we first diagnose the bottleneck of the text-to-SQL task by providing a new testbed, in which we observe that existing models present poor generalization ability on rarely-seen data. The above analysis encourages us to design a simple but effective auxiliary task, which serves as a supportive model as well as a regularization term to the generation task to increase the models' generalization. Experimentally, We evaluate our models on a large text-to-SQL dataset WikiSQL. Compared to a strong baseline coarse-to-fine model, our models improve over the baseline by more than 3% absolute in accuracy on the whole dataset. More interestingly, on a zero-shot subset test of WikiSQL, our models achieve 5% absolute accuracy gain over the baseline, clearly demonstrating its superior generalizability.
|
Shuaichen Chang, Pengfei Liu, Yun Tang, Jing Huang, Xiaodong He, Bowen Zhou
| null | null | 2,020 |
aaai
|
Planning with Abstract Learned Models While Learning Transferable Subtasks
| null |
We introduce an algorithm for model-based hierarchical reinforcement learning to acquire self-contained transition and reward models suitable for probabilistic planning at multiple levels of abstraction. We call this framework Planning with Abstract Learned Models (PALM). By representing subtasks symbolically using a new formal structure, the lifted abstract Markov decision process (L-AMDP), PALM learns models that are independent and modular. Through our experiments, we show how PALM integrates planning and execution, facilitating a rapid and efficient learning of abstract, hierarchical models. We also demonstrate the increased potential for learned models to be transferred to new and related tasks.
|
John Winder, Stephanie Milani, Matthew Landen, Erebus Oh, Shane Parr, Shawn Squire, Marie desJardins, Cynthia Matuszek
| null | null | 2,020 |
aaai
|
Neural Architecture Search Using Deep Neural Networks and Monte Carlo Tree Search
| null |
Neural Architecture Search (NAS) has shown great success in automating the design of neural networks, but the prohibitive amount of computations behind current NAS methods requires further investigations in improving the sample efficiency and the network evaluation cost to get better results in a shorter time. In this paper, we present a novel scalable Monte Carlo Tree Search (MCTS) based NAS agent, named AlphaX, to tackle these two aspects. AlphaX improves the search efficiency by adaptively balancing the exploration and exploitation at the state level, and by a Meta-Deep Neural Network (DNN) to predict network accuracies for biasing the search toward a promising region. To amortize the network evaluation cost, AlphaX accelerates MCTS rollouts with a distributed design and reduces the number of epochs in evaluating a network by transfer learning, which is guided with the tree structure in MCTS. In 12 GPU days and 1000 samples, AlphaX found an architecture that reaches 97.84% top-1 accuracy on CIFAR-10, and 75.5% top-1 accuracy on ImageNet, exceeding SOTA NAS methods in both the accuracy and sampling efficiency. Particularly, we also evaluate AlphaX on NASBench-101, a large scale NAS dataset; AlphaX is 3x and 2.8x more sample efficient than Random Search and Regularized Evolution in finding the global optimum. Finally, we show the searched architecture improves a variety of vision applications from Neural Style Transfer, to Image Captioning and Object Detection.
|
Linnan Wang, Yiyang Zhao, Yuu Jinnai, Yuandong Tian, Rodrigo Fonseca
| null | null | 2,020 |
aaai
|
Computing Superior Counter-Examples for Conformant Planning
| null |
In a counter-example based approach to conformant planning, choosing the right counter-example can improve performance. We formalise this observation by introducing the notion of “superiority” of a counter-example over another one, that holds whenever the superior counter-example exhibits more tags than the latter. We provide a theoretical explanation that supports the strategy of searching for maximally superior counter-examples, and we show how this strategy can be implemented. The empirical experiments validate our approach.
|
Xiaodi Zhang, Alban Grastien, Enrico Scala
| null | null | 2,020 |
aaai
|
Reshaping Diverse Planning
| null |
The need for multiple plans has been established by various planning applications. In some, solution quality has the predominant role, while in others diversity is the key factor. Most recent work takes both plan quality and solution diversity into account under the generic umbrella of diverse planning. There is no common agreement, however, on a collection of computational problems that fall under that generic umbrella. This in particular might lead to a comparison between planners that have different solution guarantees or optimization criteria in mind. In this work we revisit diverse planning literature in search of such a collection of computational problems, classifying the existing planners to these problems. We formally define a taxonomy of computational problems with respect to both plan quality and solution diversity, extending the existing work. We propose a novel approach to diverse planning, exploiting existing classical planners via planning task reformulation and choosing a subset of plans of required size in post-processing. Based on that, we present planners for two computational problems, that most existing planners solve. Our experiments show that the proposed approach significantly improves over the best performing existing planners in terms of coverage, the overall solution quality, and the overall diversity according to various diversity metrics.
|
Michael Katz, Shirin Sohrabi
| null | null | 2,020 |
aaai
|
Refining HTN Methods via Task Insertion with Preferences
| null |
Hierarchical Task Network (HTN) planning is showing its power in real-world planning. Although domain experts have partial hierarchical domain knowledge, it is time-consuming to specify all HTN methods, leaving them incomplete. On the other hand, traditional HTN learning approaches focus only on declarative goals, omitting the hierarchical domain knowledge. In this paper, we propose a novel learning framework to refine HTN methods via task insertion with completely preserving the original methods. As it is difficult to identify incomplete methods without designating declarative goals for compound tasks, we introduce the notion of prioritized preference to capture the incompleteness possibility of methods. Specifically, the framework first computes the preferred completion profile w.r.t. the prioritized preference to refine the incomplete methods. Then it finds the minimal set of refined methods via a method substitution operation. Experimental analysis demonstrates that our approach is effective, especially in solving new HTN planning instances.
|
Zhanhao Xiao, Hai Wan, Hankui Hankz Zhuo, Andreas Herzig, Laurent Perrussel, Peilin Chen
| null | null | 2,020 |
aaai
|
Generalized Planning with Positive and Negative Examples
| null |
Generalized planning aims at computing an algorithm-like structure (generalized plan) that solves a set of multiple planning instances. In this paper we define negative examples for generalized planning as planning instances that must not be solved by a generalized plan. With this regard the paper extends the notion of validation of a generalized plan as the problem of verifying that a given generalized plan solves the set of input positives instances while it fails to solve a given input set of negative examples. This notion of plan validation allows us to define quantitative metrics to asses the generalization capacity of generalized plans. The paper also shows how to incorporate this new notion of plan validation into a compilation for plan synthesis that takes both positive and negative instances as input. Experiments show that incorporating negative examples can accelerate plan synthesis in several domains and leverage quantitative metrics to evaluate the generalization capacity of the synthesized plans.
|
Javier Segovia-Aguas, Sergio Jiménez, Anders Jonsson
| null | null | 2,020 |
aaai
|
Idle Time Optimization for Target Assignment and Path Finding in Sortation Centers
| null |
In this paper, we study the one-shot and lifelong versions of the Target Assignment and Path Finding problem in automated sortation centers, where each agent needs to constantly assign itself a sorting station, move to its assigned station without colliding with obstacles or other agents, wait in the queue of that station to obtain a parcel for delivery, and then deliver the parcel to a sorting bin. The throughput of such centers is largely determined by the total idle time of all stations since their queues can frequently become empty. To address this problem, we first formalize and study the one-shot version that assigns stations to a set of agents and finds collision-free paths for the agents to their assigned stations. We present efficient algorithms for this task based on a novel min-cost max-flow formulation that minimizes the total idle time of all stations in a fixed time window. We then demonstrate how our algorithms for solving the one-shot problem can be applied to solving the lifelong problem as well. Experimentally, we believe to be the first researchers to consider real-world automated sortation centers using an industrial simulator with realistic data and a kinodynamic model of real robots. On this simulator, we showcase the benefits of our algorithms by demonstrating their efficiency and effectiveness for up to 350 agents.
|
Ngai Meng Kou, Cheng Peng, Hang Ma, T. K. Satish Kumar, Sven Koenig
| null | null | 2,020 |
aaai
|
HDDL: An Extension to PDDL for Expressing Hierarchical Planning Problems
| null |
The research in hierarchical planning has made considerable progress in the last few years. Many recent systems do not rely on hand-tailored advice anymore to find solutions, but are supposed to be domain-independent systems that come with sophisticated solving techniques. In principle, this development would make the comparison between systems easier (because the domains are not tailored to a single system anymore) and – much more important – also the integration into other systems, because the modeling process is less tedious (due to the lack of advice) and there is no (or less) commitment to a certain planning system the model is created for. However, these advantages are destroyed by the lack of a common input language and feature set supported by the different systems. In this paper, we propose an extension to PDDL, the description language used in non-hierarchical planning, to the needs of hierarchical planning systems.
|
Daniel Höller, Gregor Behnke, Pascal Bercher, Susanne Biundo, Humbert Fiorino, Damien Pellier, Ron Alford
| null | null | 2,020 |
aaai
|
Top-Quality Planning: Finding Practically Useful Sets of Best Plans
| null |
The need for finding a set of plans rather than one has been motivated by a variety of planning applications. The problem is studied in the context of both diverse and top-k planning: while diverse planning focuses on the difference between pairs of plans, the focus of top-k planning is on the quality of each individual plan. Recent work in diverse planning introduced additionally restrictions on solution quality. Naturally, there are application domains where diversity plays the major role and domains where quality is the predominant feature. In both cases, however, the amount of produced plans is often an artificial constraint, and therefore the actual number has little meaning. Inspired by the recent work in diverse planning, we propose a new family of computational problems called top-quality planning, where solution validity is defined through plan quality bound rather than an arbitrary number of plans. Switching to bounding plan quality allows us to implicitly represent sets of plans. In particular, it makes it possible to represent sets of plans that correspond to valid plan reorderings with a single plan. We formally define the unordered top-quality planning computational problem and present the first planner for that problem. We empirically demonstrate the superior performance of our approach compared to a top-k planner-based baseline, ranging from 41% increase in coverage for finding all optimal plans to 69% increase in coverage for finding all plans of quality up to 120% of optimal plan cost. Finally, complementing the new approach by a complete procedure for generating all valid reorderings of a given plan, we derive a top-quality planner. We show the planner to be competitive with a top-k planner based baseline.
|
Michael Katz, Shirin Sohrabi, Octavian Udrea
| null | null | 2,020 |
aaai
|
Semantic Attachments for HTN Planning
| null |
Hierarchical Task Networks (HTN) planning uses a decomposition process guided by domain knowledge to guide search towards a planning task. While many HTN planners allow calls to external processes (e.g. to a simulator interface) during the decomposition process, this is a computationally expensive process, so planner implementations often use such calls in an ad-hoc way using very specialized domain knowledge to limit the number of calls. Conversely, the classical planners that are capable of using external calls (often called semantic attachments) during planning are limited to generating a fixed number of ground operators at problem grounding time. We formalize Semantic Attachments for HTN planning using semi coroutines, allowing such procedurally defined predicates to link the planning process to custom unifications outside of the planner, such as numerical results from a robotics simulator. The resulting planner then uses such coroutines as part of its backtracking mechanism to search through parallel dimensions of the state-space (e.g. through numeric variables). We show empirically that our planner outperforms the state-of-the-art numeric planners in a number of domains using minimal extra domain knowledge.
|
Maurício Cecílio Magnaguagno, Felipe Meneguzzi
| null | null | 2,020 |
aaai
|
Decidability and Complexity of Action-Based Temporal Planning over Dense Time
| null |
This paper studies the computational complexity of temporal planning, as represented by PDDL 2.1, interpreted over dense time. When time is considered discrete, the problem is known to be EXPSPACE-complete. However, the official PDDL 2.1 semantics, and many implementations, interpret time as a dense domain. This work provides several results about the complexity of the problem, studying a few interesting cases: whether a minimum amount ϵ of separation between mutually exclusive events is given, in contrast to the separation being simply required to be non-zero, and whether or not actions are allowed to overlap already running instances of themselves. We prove the problem to be PSPACE-complete when self-overlap is forbidden, whereas, when allowed, it becomes EXPSPACE-complete with ϵ-separation and undecidable with non-zero separation. These results clarify the computational consequences of different choices in the definition of the PDDL 2.1 semantics, which were vague until now.
|
Nicola Gigante, Andrea Micheli, Angelo Montanari, Enrico Scala
| null | null | 2,020 |
aaai
|
Monte Carlo Tree Search in Continuous Spaces Using Voronoi Optimistic Optimization with Regret Bounds
| null |
Many important applications, including robotics, data-center management, and process control, require planning action sequences in domains with continuous state and action spaces and discontinuous objective functions. Monte Carlo tree search (MCTS) is an effective strategy for planning in discrete action spaces. We provide a novel MCTS algorithm (voot) for deterministic environments with continuous action spaces, which, in turn, is based on a novel black-box function-optimization algorithm (voo) to efficiently sample actions. The voo algorithm uses Voronoi partitioning to guide sampling, and is particularly efficient in high-dimensional spaces. The voot algorithm has an instance of voo at each node in the tree. We provide regret bounds for both algorithms and demonstrate their empirical effectiveness in several high-dimensional problems including two difficult robotics planning problems.
|
Beomjoon Kim, Kyungjae Lee, Sungbin Lim, Leslie Kaelbling, Tomas Lozano-Perez
| null | null | 2,020 |
aaai
|
Solving Sum-of-Costs Multi-Agent Pathfinding with Answer-Set Programming
| null |
Solving a Multi-Agent Pathfinding (MAPF) problem involves finding non-conflicting paths that lead a number of agents to their goal location. In the sum-of-costs variant of MAPF, one is also required to minimize the total number of moves performed by agents before stopping at the goal. Not surprisingly, since MAPF is combinatorial, a number of compilations to Satisfiability solving (SAT) and Answer Set Programming (ASP) exist. In this paper, we propose the first family of compilations to ASP that solve sum-of-costs MAPF over 4-connected grids. Unlike existing compilations to ASP that we are aware of, our encoding is the first that, after grounding, produces a number of clauses that is linear on the number of agents. In addition, the representation of the optimization objective is also carefully written, such that its size after grounding does not depend on the size of the grid. In our experimental evaluation, we show that our approach outperforms search- and SAT-based sum-of-costs MAPF solvers when grids are congested with agents.
|
Rodrigo N. Gómez, Carlos Hernández, Jorge A. Baier
| null | null | 2,020 |
aaai
|
Dynamic Control of Probabilistic Simple Temporal Networks
| null |
The controllability of a temporal network is defined as an agent's ability to navigate around the uncertainty in its schedule and is well-studied for certain networks of temporal constraints. However, many interesting real-world problems can be better represented as Probabilistic Simple Temporal Networks (PSTNs) in which the uncertain durations are represented using potentially-unbounded probability density functions. This can make it inherently impossible to control for all eventualities. In this paper, we propose two new dynamic controllability algorithms that attempt to maximize the likelihood of successfully executing a schedule within a PSTN. The first approach, which we call Min-Loss DC, finds a dynamic scheduling strategy that minimizes loss of control by using a conflict-directed search to decide where to sacrifice the control in a way that optimizes overall success. The second approach, which we call Max-Gain DC, works in the other direction: it finds a dynamically controllable schedule and then attempts to progressively strengthen it by capturing additional uncertainty. Our approaches are the first known that work by finding maximally dynamically controllable schedules. We empirically compare our approaches against two existing PSTN offline dispatch approaches and one online approach and show that our Min-Loss DC algorithm outperforms the others in terms of maximizing execution success while maintaining competitive runtimes.
|
Michael Gao, Lindsay Popowski, Jim Boerkoel
| null | null | 2,020 |
aaai
|
Temporal Planning with Intermediate Conditions and Effects
| null |
Automated temporal planning is the technology of choice when controlling systems that can execute more actions in parallel and when temporal constraints, such as deadlines, are needed in the model. One limitation of several action-based planning systems is that actions are modeled as intervals having conditions and effects only at the extremes and as invariants, but no conditions nor effects can be specified at arbitrary points or sub-intervals.In this paper, we address this limitation by providing an effective heuristic-search technique for temporal planning, allowing the definition of actions with conditions and effects at any arbitrary time within the action duration. We experimentally demonstrate that our approach is far better than standard encodings in PDDL 2.1 and is competitive with other approaches that can (directly or indirectly) represent intermediate action conditions or effects.
|
Alessandro Valentini, Andrea Micheli, Alessandro Cimatti
| null | null | 2,020 |
aaai
|
On Succinct Groundings of HTN Planning Problems
| null |
Both search-based and translation-based planning systems usually operate on grounded representations of the problem. Planning models, however, are commonly defined using lifted description languages. Thus, planning systems usually generate a grounded representation of the lifted model as a preprocessing step. For HTN planning models, only one method to ground lifted models has been published so far. In this paper we present a new approach for grounding HTN planning problems that produces smaller groundings in a shorter timespan than the previously published method.
|
Gregor Behnke, Daniel Höller, Alexander Schmid, Pascal Bercher, Susanne Biundo
| null | null | 2,020 |
aaai
|
Automated Synthesis of Social Laws in STRIPS
| null |
Agents operating in a multi-agent environment must consider not just their actions, but also those of the other agents in the system. Artificial social systems are a well-known means for coordinating a set of agents, without requiring centralized planning or online negotiation between agents. Artificial social systems enact a social law which restricts the agents from performing some actions under some circumstances. A robust social law prevents the agents from interfering with each other, but does not prevent them from achieving their goals. Previous work has addressed how to check if a given social law, formulated in a variant of ma-strips, is robust, via compilation to planning. However, the social law was manually specified. In this paper, we address the problem of automatically synthesizing a robust social law for a given multi-agent environment. We treat the problem of social law synthesis as a search through the space of possible social laws, relying on the robustness verification procedure as a goal test. We also show how to exploit additional information produced by the robustness verification procedure to guide the search.
|
Ronen Nir, Alexander Shleyfman, Erez Karpas
| null | null | 2,020 |
aaai
|
Planning and Acting with Non-Deterministic Events: Navigating between Safe States
| null |
Automated Planning addresses the problem of finding a sequence of actions, a plan, transforming the environment from its initial state to some goal state. In real-world environments, exogenous events might occur and might modify the environment without agent's consent. Besides disrupting agent's plan, events might hinder agent's pursuit towards its goals and even cause damage (e.g. destroying the robot).In this paper, we leverage the notion of Safe States in dynamic environments under presence of non-deterministic exogenous events that might eventually cause dead-ends (e.g. “damage” the agent) if the agent is not careful while executing its plan. We introduce a technique for generating plans that constrains the number of consecutive “unsafe” actions in a plan and a technique for generating “robust” plans that effectively evade event effects. Combination of both approaches plans and executes robust plans between safe states. We empirically show that such an approach effectively navigates the agent towards its goals in spite of presence of dead-ends.
|
Lukas Chrpa, Jakub Gemrot, Martin Pilat
| null | null | 2,020 |
aaai
|
A New Approach to Plan-Space Explanation: Analyzing Plan-Property Dependencies in Oversubscription Planning
| null |
In many usage scenarios of AI Planning technology, users will want not just a plan π but an explanation of the space of possible plans, justifying π. In particular, in oversubscription planning where not all goals can be achieved, users may ask why a conjunction A of goals is not achieved by π. We propose to answer this kind of question with the goal conjunctions B excluded by A, i. e., that could not be achieved if A were to be enforced. We formalize this approach in terms of plan-property dependencies, where plan properties are propositional formulas over the goals achieved by a plan, and dependencies are entailment relations in plan space. We focus on entailment relations of the form ∧g∈Ag ⇒ ⌝ ∧g∈Bg, and devise analysis techniques globally identifying all such relations, or locally identifying the implications of a single given plan property (user question) ∧g∈Ag. We show how, via compilation, one can analyze dependencies between a richer form of plan properties, specifying formulas over action subsets touched by the plan. We run comprehensive experiments on adapted IPC benchmarks, and find that the suggested analyses are reasonably feasible at the global level, and become significantly more effective at the local level.
|
Rebecca Eifler, Michael Cashmore, Jörg Hoffmann, Daniele Magazzeni, Marcel Steinmetz
| null | null | 2,020 |
aaai
|
Beliefs We Can Believe in: Replacing Assumptions with Data in Real-Time Search
| null |
Suboptimal heuristic search algorithms can benefit from reasoning about heuristic error, especially in a real-time setting where there is not enough time to search all the way to a goal. However, current reasoning methods implicitly or explicitly incorporate assumptions about the cost-to-go function. We consider a recent real-time search algorithm, called Nancy, that manipulates explicit beliefs about the cost-to-go. The original presentation of Nancy assumed that these beliefs are Gaussian, with parameters following a certain form. In this paper, we explore how to replace these assumptions with actual data. We develop a data-driven variant of Nancy, DDNancy, that bases its beliefs on heuristic performance statistics from the same domain. We extend Nancy and DDNancy with the notion of persistence and prove their completeness. Experimental results show that DDNancy can perform well in domains in which the original assumption-based Nancy performs poorly.
|
Maximilian Fickert, Tianyi Gu, Leonhard Staut, Wheeler Ruml, Joerg Hoffmann, Marek Petrik
| null | null | 2,020 |
aaai
|
Optimizing Reachability Sets in Temporal Graphs by Delaying
| null |
A temporal graph is a dynamic graph where every edge is assigned a set of integer time labels that indicate at which discrete time step the edge is available. In this paper, we study how changes of the time labels, corresponding to delays on the availability of the edges, affect the reachability sets from given sources. The questions about reachability sets are motivated by numerous applications of temporal graphs in network epidemiology and scheduling problems in supply networks in manufacturing. We introduce control mechanisms for reachability sets that are based on two natural operations of delaying time events. The first operation, termed merging, is global and batches together consecutive time labels in the whole network simultaneously. This corresponds to postponing all events until a particular time. The second, imposes independent delays on the time labels of every edge of the graph. We provide a thorough investigation of the computational complexity of different objectives related to reachability sets when these operations are used. For the merging operation, we prove NP-hardness results for several minimization and maximization reachability objectives, even for very simple graph structures. For the second operation, we prove that the minimization problems are NP-hard when the number of allowed delays is bounded. We complement this with a polynomial-time algorithm for the case of unbounded delays.
|
Argyrios Deligkas, Igor Potapov
| null | null | 2,020 |
aaai
|
Novel Is Not Always Better: On the Relation between Novelty and Dominance Pruning
| null |
Novelty pruning is a planning technique that focuses on exploring states that are novel, i.e., those containing facts that have not been seen before. This seemingly simple idea has had a huge impact on the state of the art in planning though its effectiveness is not entirely understood yet.We relate novelty to dominance pruning, which compares states to previously seen states to eliminate those that are provably worse in terms of goal distance. Novelty can be interpreted as an unsafe approximation of dominance, where states containing novel facts are relevant because they enable new paths to the goal and, therefore, they are less likely to be dominated by others. This provides a framework to understand the success of novelty, resulting in new variants that combine both techniques.
|
Joschka Gross, Alvaro Torralba, Maximilian Fickert
| null | null | 2,020 |
aaai
|
Time-Inconsistent Planning: Simple Motivation Is Hard to Find
| null |
People sometimes act differently when making decisions affecting the present moment versus decisions affecting the future only. This is referred to as time-inconsistent behaviour, and can be modeled as agents exhibiting present bias. A resulting phenomenon is abandonment, which is when an agent initially pursues a task, but ultimately gives up before reaping the rewards. With the introduction of the graph-theoretic time-inconsistent planning model due to Kleinberg and Oren, it has been possible to investigate the computational complexity of how a task designer best can support a present-biased agent in completing the task. In this paper, we study the complexity of finding a choice reduction for the agent; that is, how to remove edges and vertices from the task graph such that a present-biased agent will remain motivated to reach his target even for a limited reward. While this problem is NP-complete in general, this is not necessarily true for instances which occur in practice, or for solutions which are of interest to task designers. For instance, a task designer may desire to find the best task graph which is not too complicated. We therefore investigate the problem of finding simple motivating subgraphs. These are structures where the agent will modify his plan at most k times along the way. We quantify this simplicity in the time-inconsistency model as a structural parameter: The number of branching vertices (vertices with out-degree at least 2) in a minimal motivating subgraph. Our results are as follows: We give a linear algorithm for finding an optimal motivating path, i. e. when k = 0. On the negative side, we show that finding a simple motivating subgraph is NP-complete even if we allow only a single branching vertex — revealing that simple motivating subgraphs are indeed hard to find. However, we give a pseudo-polynomial algorithm for the case when k is fixed and edge weights are rationals, which might be a reasonable assumption in practice.
|
Fedor V. Fomin, Torstein J. F. Strømme
| null | null | 2,020 |
aaai
|
A New Framework for Online Testing of Heterogeneous Treatment Effect
| null |
We propose a new framework for online testing of heterogeneous treatment effects. The proposed test, named sequential score test (SST), is able to control type I error under continuous monitoring and detect multi-dimensional heterogeneous treatment effects. We provide an online p-value calculation for SST, making it convenient for continuous monitoring, and extend our tests to online multiple testing settings by controlling the false discovery rate. We examine the empirical performance of the proposed tests and compare them with a state-of-art online test, named mSPRT using simulations and a real data. The results show that our proposed test controls type I error at any time, has higher detection power and allows quick inference on online A/B testing.
|
Miao Yu, Wenbin Lu, Rui Song
| null | null | 2,020 |
aaai
|
Reinforcement Learning of Risk-Constrained Policies in Markov Decision Processes
| null |
Markov decision processes (MDPs) are the defacto framework for sequential decision making in the presence of stochastic uncertainty. A classical optimization criterion for MDPs is to maximize the expected discounted-sum payoff, which ignores low probability catastrophic events with highly negative impact on the system. On the other hand, risk-averse policies require the probability of undesirable events to be below a given threshold, but they do not account for optimization of the expected payoff. We consider MDPs with discounted-sum payoff with failure states which represent catastrophic outcomes. The objective of risk-constrained planning is to maximize the expected discounted-sum payoff among risk-averse policies that ensure the probability to encounter a failure state is below a desired threshold. Our main contribution is an efficient risk-constrained planning algorithm that combines UCT-like search with a predictor learned through interaction with the MDP (in the style of AlphaZero) and with a risk-constrained action selection via linear programming. We demonstrate the effectiveness of our approach with experiments on classical MDPs from the literature, including benchmarks with an order of 106 states.
|
Tomáš Brázdil, Krishnendu Chatterjee, Petr Novotný, Jiří Vahala
| null | null | 2,020 |
aaai
|
Lifted Fact-Alternating Mutex Groups and Pruned Grounding of Classical Planning Problems
| null |
In this paper, we focus on the inference of mutex groups in the lifted (PDDL) representation. We formalize the inference and prove that the most commonly used translator from the Fast Downward (FD) planning system infers a certain subclass of mutex groups, called fact-alternating mutex groups (fam-groups). Based on that, we show that the previously proposed fam-groups-based pruning techniques for the STRIPS representation can be utilized during the grounding process with lifted fam-groups, i.e., before the full STRIPS representation is known. Furthermore, we propose an improved inference algorithm for lifted fam-groups that produces a richer set of fam-groups than the FD translator and we demonstrate a positive impact on the number of pruned operators and overall coverage.
|
Daniel Fišer
| null | null | 2,020 |
aaai
|
A Simultaneous Discover-Identify Approach to Causal Inference in Linear Models
| null |
Modern causal analysis involves two major tasks, discovery and identification. The first aims to learn a causal structure compatible with the available data, the second leverages that structure to estimate causal effects. Rather than performing the two tasks in tandem, as is usually done in the literature, we propose a symbiotic approach in which the two are performed simultaneously for mutual benefit; information gained through identification helps causal discovery and vice versa. This approach enables the usage of Verma constraints, which remain dormant in constraint-based methods of discovery, and permit us to learn more complete structures, hence identify a larger set of causal effects than previously achievable with standard methods.
|
Chi Zhang, Bryant Chen, Judea Pearl
| null | null | 2,020 |
aaai
|
Modeling Probabilistic Commitments for Maintenance Is Inherently Harder than for Achievement
| null |
Most research on probabilistic commitments focuses on commitments to achieve enabling preconditions for other agents. Our work reveals that probabilistic commitments to instead maintain preconditions for others are surprisingly harder to use well than their achievement counterparts, despite strong semantic similarities. We isolate the key difference as being not in how the commitment provider is constrained, but rather in how the commitment recipient can locally use the commitment specification to approximately model the provider's effects on the preconditions of interest. Our theoretic analyses show that we can more tightly bound the potential suboptimality due to approximate modeling for achievement than for maintenance commitments. We empirically evaluate alternative approximate modeling strategies, confirming that probabilistic maintenance commitments are qualitatively more challenging for the recipient to model well, and indicating the need for more detailed specifications that can sacrifice some of the agents' autonomy.
|
Qi Zhang, Edmund Durfee, Satinder Singh
| null | null | 2,020 |
aaai
|
Hybrid Compositional Reasoning for Reactive Synthesis from Finite-Horizon Specifications
| null |
LTLf synthesis is the automated construction of a reactive system from a high-level description, expressed in LTLf, of its finite-horizon behavior. So far, the conversion of LTLf formulas to deterministic finite-state automata (DFAs) has been identified as the primary bottleneck to the scalabity of synthesis. Recent investigations have also shown that the size of the DFA state space plays a critical role in synthesis as well.Therefore, effective resolution of the bottleneck for synthesis requires the conversion to be time and memory performant, and prevent state-space explosion. Current conversion approaches, however, which are based either on explicit-state representation or symbolic-state representation, fail to address these necessities adequately at scale: Explicit-state approaches generate minimal DFA but are slow due to expensive DFA minimization. Symbolic-state representations can be succinct, but due to the lack of DFA minimization they generate such large state spaces that even their symbolic representations cannot compensate for the blow-up.This work proposes a hybrid representation approach for the conversion. Our approach utilizes both explicit and symbolic representations of the state-space, and effectively leverages their complementary strengths. In doing so, we offer an LTLf to DFA conversion technique that addresses all three necessities, hence resolving the bottleneck. A comprehensive empirical evaluation on conversion and synthesis benchmarks supports the merits of our hybrid approach.
|
Suguman Bansal, Yong Li, Lucas Tabajara, Moshe Vardi
| null | null | 2,020 |
aaai
|
Information Shaping for Enhanced Goal Recognition of Partially-Informed Agents
| null |
We extend goal recognition design to account for partially informed agents. In particular, we consider a two-agent setting in which one agent, the actor, seeks to achieve a goal but has only incomplete information about the environment. The second agent, the recognizer, has perfect information and aims to recognize the actor's goal from its behavior as quickly as possible. As a one-time offline intervention and with the objective of facilitating the recognition task, the recognizer can selectively reveal information to the actor. The problem of selecting which information to reveal, which we call information shaping, is challenging not only because the space of information shaping options may be large, but also because more information revelation need not make it easier to recognize an agent's goal. We formally define this problem, and suggest a pruning approach for efficiently searching the search space. We demonstrate the effectiveness and efficiency of the suggested method on standard benchmarks.
|
Sarah Keren, Haifeng Xu, Kofi Kwapong, David Parkes, Barbara Grosz
| null | null | 2,020 |
aaai
|
POP ≡ POCL, Right? Complexity Results for Partial Order (Causal Link) Makespan Minimization
| null |
We study PO and POCL plans with regard to their makespan – the execution time when allowing the parallel execution of causally independent actions. Partially ordered (PO) plans are often assumed to be equivalent to partial order causal link (POCL) plans, where the causal relationships between actions are explicitly represented via causal links. As a first contribution, we study the similarities and differences of PO and POCL plans, thereby clarifying a common misconception about their relationship: There are PO plans for which there does not exist a POCL plan with the same orderings. We prove that we can still always find a POCL plan with the same makespan in polynomial time. As another main result we prove that turning a PO or POCL plan into one with minimal makespan by only removing ordering constraints (called deordering) is NP-complete. We provide a series of further results on special cases and implications, such as reordering, where orderings can be changed arbitrarily.
|
Pascal Bercher, Conny Olz
| null | null | 2,020 |
aaai
|
Parallel AND/OR Search for Marginal MAP
| null |
Marginal MAP is a difficult mixed inference task for graphical models. Existing state-of-the-art algorithms for solving exactly this task are based on either depth-first or best-first sequential search over an AND/OR search space. In this paper, we explore and evaluate for the first time the power of parallel search for exact Marginal MAP inference. We introduce a new parallel shared-memory recursive best-first AND/OR search algorithm that explores the search space in a best-first manner while operating with limited memory. Subsequently, we develop a complete parallel search scheme that only parallelizes the conditional likelihood computations. We also extend the proposed algorithms into depth-first parallel search schemes. Our experiments on difficult benchmarks demonstrate the effectiveness of the parallel search algorithms against current sequential methods for solving Marginal MAP exactly.
|
Radu Marinescu, Akihiro Kishimoto, Adi Botea
| null | null | 2,020 |
aaai
|
Error-Correcting and Verifiable Parallel Inference in Graphical Models
| null |
We present a novel framework for parallel exact inference in graphical models. Our framework supports error-correction during inference and enables fast verification that the result of inference is correct, with probabilistic soundness. The computational complexity of inference essentially matches the cost of w-cutset conditioning, a known generalization of Pearl's classical loop-cutset conditioning for inference. Verifying the result for correctness can be done with as little as essentially the square root of the cost of inference. Our main technical contribution amounts to designing a low-degree polynomial extension of the cutset approach, and then reducing to a univariate polynomial employing techniques recently developed for noninteractive probabilistic proof systems.
|
Negin Karimi, Petteri Kaski, Mikko Koivisto
| null | null | 2,020 |
aaai
|
Temporal Logics Over Finite Traces with Uncertainty
| null |
Temporal logics over finite traces have recently seen wide application in a number of areas, from business process modelling, monitoring, and mining to planning and decision making. However, real-life dynamic systems contain a degree of uncertainty which cannot be handled with classical logics. We thus propose a new probabilistic temporal logic over finite traces using superposition semantics, where all possible evolutions are possible, until observed. We study the properties of the logic and provide automata-based mechanisms for deriving probabilistic inferences from its formulas. We then study a fragment of the logic with better computational properties. Notably, formulas in this fragment can be discovered from event log data using off-the-shelf existing declarative process discovery techniques.
|
Fabrizio M Maggi, Marco Montali, Rafael Peñaloza
| null | null | 2,020 |
aaai
|
Beyond the Grounding Bottleneck: Datalog Techniques for Inference in Probabilistic Logic Programs
| null |
State-of-the-art inference approaches in probabilistic logic programming typically start by computing the relevant ground program with respect to the queries of interest, and then use this program for probabilistic inference using knowledge compilation and weighted model counting. We propose an alternative approach that uses efficient Datalog techniques to integrate knowledge compilation with forward reasoning with a non-ground program. This effectively eliminates the grounding bottleneck that so far has prohibited the application of probabilistic logic programming in query answering scenarios over knowledge graphs, while also providing fast approximations on classical benchmarks in the field.
|
Efthymia Tsamoura, Victor Gutierrez-Basulto, Angelika Kimmig
| null | null | 2,020 |
aaai
|
Gradient-Based Optimization for Bayesian Preference Elicitation
| null |
Effective techniques for eliciting user preferences have taken on added importance as recommender systems (RSs) become increasingly interactive and conversational. A common and conceptually appealing Bayesian criterion for selecting queries is expected value of information (EVOI). Unfortunately, it is computationally prohibitive to construct queries with maximum EVOI in RSs with large item spaces. We tackle this issue by introducing a continuous formulation of EVOI as a differentiable network that can be optimized using gradient methods available in modern machine learning computational frameworks (e.g., TensorFlow, PyTorch). We exploit this to develop a novel Monte Carlo method for EVOI optimization, which is much more scalable for large item spaces than methods requiring explicit enumeration of items. While we emphasize the use of this approach for pairwise (or k-wise) comparisons of items, we also demonstrate how our method can be adapted to queries involving subsets of item attributes or “partial items,” which are often more cognitively manageable for users. Experiments show that our gradient-based EVOI technique achieves state-of-the-art performance across several domains while scaling to large item spaces.
|
Ivan Vendrov, Tyler Lu, Qingqing Huang, Craig Boutilier
| null | null | 2,020 |
aaai
|
Factorized Inference in Deep Markov Models for Incomplete Multimodal Time Series
| null |
Integrating deep learning with latent state space models has the potential to yield temporal models that are powerful, yet tractable and interpretable. Unfortunately, current models are not designed to handle missing data or multiple data modalities, which are both prevalent in real-world data. In this work, we introduce a factorized inference method for Multimodal Deep Markov Models (MDMMs), allowing us to filter and smooth in the presence of missing data, while also performing uncertainty-aware multimodal fusion. We derive this method by factorizing the posterior p(z|x) for non-linear state space models, and develop a variational backward-forward algorithm for inference. Because our method handles incompleteness over both time and modalities, it is capable of interpolation, extrapolation, conditional generation, label prediction, and weakly supervised learning of multimodal time series. We demonstrate these capabilities on both synthetic and real-world multimodal data under high levels of data deletion. Our method performs well even with more than 50% missing data, and outperforms existing deep approaches to inference in latent time series.
|
Tan Zhi-Xuan, Harold Soh, Desmond Ong
| null | null | 2,020 |
aaai
|
Estimating Causal Effects Using Weighting-Based Estimators
| null |
Causal effect identification is one of the most prominent and well-understood problems in causal inference. Despite the generality and power of the results developed so far, there are still challenges in their applicability to practical settings, arguably due to the finitude of the samples. Simply put, there is a gap between causal effect identification and estimation. One popular setting in which sample-efficient estimators from finite samples exist is when the celebrated back-door condition holds. In this paper, we extend weighting-based methods developed for the back-door case to more general settings, and develop novel machinery for estimating causal effects using the weighting-based method as a building block. We derive graphical criteria under which causal effects can be estimated using this new machinery and demonstrate the effectiveness of the proposed method through simulation studies.
|
Yonghan Jung, Jin Tian, Elias Bareinboim
| null | null | 2,020 |
aaai
|
The Choice Function Framework for Online Policy Improvement
| null |
There are notable examples of online search improving over hand-coded or learned policies (e.g. AlphaZero) for sequential decision making. It is not clear, however, whether or not policy improvement is guaranteed for many of these approaches, even when given a perfect leaf evaluation function and transition model. Indeed, simple counterexamples show that seemingly reasonable online search procedures can hurt performance compared to the original policy. To address this issue, we introduce the choice function framework for analyzing online search procedures for policy improvement. A choice function specifies the actions to be considered at every node of a search tree, with all other actions being pruned. Our main contribution is to give sufficient conditions for stationary and non-stationary choice functions to guarantee that the value achieved by online search is no worse than the original policy. In addition, we describe a general parametric class of choice functions that satisfy those conditions and present an illustrative use case of the empirical utility of the framework.
|
Murugeswari Issakkimuthu, Alan Fern, Prasad Tadepalli
| null | null | 2,020 |
aaai
|
Tandem Inference: An Out-of-Core Streaming Algorithm for Very Large-Scale Relational Inference
| null |
Statistical relational learning (SRL) frameworks allow users to create large, complex graphical models using a compact, rule-based representation. However, these models can quickly become prohibitively large and not fit into machine memory. In this work we address this issue by introducing a novel technique called tandem inference (ti). The primary idea of ti is to combine grounding and inference such that both processes happen in tandem. ti uses an out-of-core streaming approach to overcome memory limitations. Even when memory is not an issue, we show that our proposed approach is able to do inference faster while using less memory than existing approaches. To show the effectiveness of ti, we use a popular SRL framework called Probabilistic Soft Logic (PSL). We implement ti for PSL by proposing a gradient-based inference engine and a streaming approach to grounding. We show that we are able to run an SRL model with over 1B cliques in under nine hours and using only 10 GB of RAM; previous approaches required more than 800 GB for this model and are infeasible on common hardware. To the best of our knowledge, this is the largest SRL model ever run.
|
Sriram Srinivasan, Eriq Augustine, Lise Getoor
| null | null | 2,020 |
aaai
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.