arxiv_id
stringlengths 9
12
| abstract
stringlengths 402
13.8k
|
|---|---|
2401.00165
|
In open-domain Question Answering (QA) , dense retrieval is crucial for finding relevant passages for answer generation. Typically, contrastive learning is used to train a retrieval model that maps passages and queries to the same semantic space. The objective is to make similar ones closer and dissimilar ones further apart. However, training such a system is challenging due to the false negative issue, where relevant passages may be missed during data annotation. Hard negative sampling, which is commonly used to improve contrastive learning, can introduce more noise in training. This is because hard negatives are those closer to a given query, and thus more likely to be false negatives. To address this issue, we propose a novel contrastive confidence regularizer for Noise Contrastive Estimation (NCE) loss, a commonly used loss for dense retrieval. Our analysis shows that the regularizer helps dense retrieval models be more robust against false negatives with a theoretical guarantee. Additionally, we propose a model-agnostic method to filter out noisy negative passages in the dataset, improving any downstream dense retrieval models. Through experiments on three datasets, we demonstrate that our method achieves better retrieval performance in comparison to existing state-of-the-art dense retrieval systems.
|
2401.00368
|
In this paper,
we introduce a novel and simple method for obtaining high-quality text embeddings
using only synthetic data and less thank training steps.
Unlike existing methods that often depend on multi-stage intermediate pre-training
with billions of weakly-supervised text pairs,
followed by fine-tuning with a few labeled datasets,
our method does not require building complex training pipelines
or relying on manually collected datasets that are often constrained by task diversity and language coverage.
We leverage proprietary LLMs to generate diverse synthetic data
for hundreds of thousands of text embedding tasks across nearlylanguages.
We then fine-tune open-source decoder-only LLMs on the synthetic data using standard contrastive loss.
Experiments demonstrate that our method achieves strong performance on highly competitive text embedding benchmarks
without using any labeled data.
Furthermore, when fine-tuned with a mixture of synthetic and labeled data,
our model sets new state-of-the-art results on the BEIR and MTEB benchmarks.
|
2401.00448
|
Large language model (LLM) scaling laws are empirical formulas that estimate changes in model quality as a result of increasing parameter count and training data. However, these formulas, including the popular DeepMind Chinchilla scaling laws, neglect to include the cost of inference. We modify the Chinchilla scaling laws to calculate the optimal LLM parameter count and pre-training data size to train and deploy a model of a given quality and inference demand. We conduct our analysis both in terms of a compute budget and real-world costs and find that LLM researchers expecting reasonably large inference demand (~1B requests) should train models smaller and longer than Chinchilla-optimal.
|
2312.17482
|
Although BERT-style encoder models are heavily used in NLP research, many researchers do not pretrain their own BERTs from scratch due to the high cost of training.
In the past half-decade since BERT first rose to prominence, many advances have been made with other transformer architectures and training configurations that have yet to be systematically incorporated into BERT.
Here, we introduce MosaicBERT, a BERT-style encoder architecture and training recipe that is empirically optimized for fast pretraining.
This efficient architecture incorporates FlashAttention, Attention with Linear Biases (ALiBi) , Gated Linear Units (GLU) , a module to dynamically remove padded tokens, and low precision LayerNorm into the classic transformer encoder block.
The training recipe includes a 30% masking ratio for the Masked Language Modeling (MLM) objective, bfloat16 precision, and vocabulary size optimized for GPU throughput, in addition to best-practices from RoBERTa and other encoder models.
When pretrained from scratch on the C4 dataset, this base model achieves a downstream average GLUE (dev) score of 79.6 in 1.13 hours on 8 A100 80 GB GPUs at a cost of roughly $20.
We plot extensive accuracy vs. pretraining speed Pareto curves and show that MosaicBERT base and large are consistently Pareto optimal when compared to a competitive BERT base and large.
This empirical speed up in pretraining enables researchers and engineers to pretrain custom BERT-style models at low cost instead of finetune on existing generic models.
We open source our model weights and code.
|
2203.15556
|
We investigate the optimal model size and number of tokens for training a transformer language model under a given compute budget.
We find that current large language models are significantly undertrained, a consequence of the recent focus on scaling language models whilst keeping the amount of training data constant.
By training over 400 language models ranging from 70 million to over 16 billion parameters on 5 to 500 billion tokens, we find that for compute-optimal training, the model size and the number of training tokens should be scaled equally: for every doubling of model size the number of training tokens should also be doubled.
We test this hypothesis by training a predicted compute-optimal model,Chinchilla, that uses the same compute budget asGopherbut with 70B parameters and 4more more data.Chinchillauniformly and significantly outperformsGopher(280B) , GPT-3 (175B) , Jurassic-1 (178B) , and Megatron-Turing NLG (530B) on a large range of downstream evaluation tasks.
This also means thatChinchillauses substantially less compute for fine-tuning and inference, greatly facilitating downstream usage.
As a highlight,Chinchillareaches a state-of-the-art average accuracy of 67.5% on the MMLU benchmark, greater than a 7% improvement overGopher.
|
2309.02390
|
One of the most surprising puzzles in neural network generalisation isgrokking: a network with perfect training accuracy but poor generalisation will, upon further training, transition to perfect generalisation. We propose that grokking occurs when the task admits a generalising solution and a memorising solution, where the generalising solution is slower to learn but moreefficient, producing larger logits with the same parameter norm. We hypothesise that memorising circuits become more inefficient with larger training datasets while generalising circuits do not, suggesting there is a critical dataset size at which memorisation and generalisation are equally efficient. We make and confirm four novel predictions about grokking, providing significant evidence in favour of our explanation. Most strikingly, we demonstrate two novel and surprising behaviours:ungrokking, in which a network regresses from perfect to low test accuracy, andsemi-grokking, in which a network shows delayed generalisation to partial rather than perfect test accuracy.
|
2101.00436v3
|
Multi-hop reasoning (i.e., reasoning across two or more documents) is a key ingredient for NLP models that leverage large corpora to exhibit broad knowledge. To retrieve evidence passages, multi-hop models must contend with a fast-growing search space across the hops, represent complex queries that combine multiple information needs, and resolve ambiguity about the best order in which to hop between training passages. We tackle these problems via Baleen, a system that improves the accuracy of multi-hop retrieval while learning robustly from weak training signals in the many-hop setting. To tame the search space, we proposecondensed retrieval, a pipeline that summarizes the retrieved passages after each hop into a single compact context. To model complex queries, we introduce afocused late interactionretriever that allows different parts of the same query representation to match disparate relevant passages. Lastly, to infer the hopping dependencies among unordered training passages, we deviselatent hop ordering, a weak-supervision strategy in which the trained retriever itself selects the sequence of hops. We evaluate Baleen on retrieval for two-hop question answering and many-hop claim verification, establishing state-of-the-art performance.
|
2305.17066v1
|
Both Minsky’s “society of mind” and Schmidhuber’s “learning to think” inspire diverse societies of large multimodal neural networks (NNs) that solve problems by interviewing each other in a “mindstorm.”
Recent implementations of NN-based societies of minds consist of large language models (LLMs) and other NN-based experts communicating through a natural language interface.
In doing so, they overcome the limitations of single LLMs, improving multimodal zero-shot reasoning.
In these natural language-based societies of mind (NLSOMs) , new agents—all communicating through the same universal symbolic language—are easily added in a modular fashion.
To demonstrate the power of NLSOMs, we assemble and experiment with several of them (having up to 129 members) , leveraging mindstorms in them to solve some practical AI tasks: visual question answering, image captioning, text-to-image synthesis, 3D generation, egocentric retrieval, embodied AI, and general language-based task solving.
We view this as a starting point towards much larger NLSOMs with billions of agents—some of which may be humans.
And with this emergence of great societies of heterogeneous minds, many new research questions have suddenly become paramount to the future of artificial intelligence.
What should be the social structure of an NLSOM?
What would be the (dis) advantages of having a monarchical rather than a democratic structure?
How can principles of NN economies be used to maximize the total reward of a reinforcement learning NLSOM?
In this work, we identify, discuss, and try to answer some of these questions.
|
2312.16682v1
|
Practitioners commonly align large language models using pairwise preferences, i.e., given labels of the type response A is preferred to response B for a given input.
Perhaps less commonly, methods have also been developed for binary feedback,
i.e. training models given labels of type response A is good or bad. We show how an existing performant binary feedback method, the Cringe Loss, can be generalized
to the pairwise preference setting using a simple soft margin extension.
Pairwise Cringe Loss is straightforward to
implement and efficient to train, and we find it outperforms state-of-the-art preference optimization algorithms such as PPO and DPO on the AlpacaFarm benchmark.
|
2112.00114
|
Large pre-trained language models perform remarkably well on tasks that can be done “in one pass”, such as generating realistic textor synthesizing computer programs.
However, they struggle with tasks that require unbounded multi-step computation, such as adding integersorexecutingprograms.
Surprisingly, we find that these same models are able to perform complex multi-step computations—even in the few-shot regime—when asked to perform the operation “step by step”, showing the results of intermediate computations.
In particular, we train Transformers to perform multi-step computations by asking them to emit intermediate computation steps into a “scratchpad”.
On a series of increasingly complex tasks ranging from long addition to the execution of arbitrary programs, we show that scratchpads dramatically improve the ability of language models to perform multi-step computations.
|
2102.12459
|
Large language models have become increasingly difficult to train because of the growing computation time and cost.
In this work, we present SRU++, a highly-efficient architecture that combines fast recurrence and attention for sequence modeling.
SRU++ exhibits strong modeling capacity and training efficiency.
On standard language modeling tasks such asenwik8,Wiki-103andbillion worddatasets, our model obtains better bits-per-character and perplexity while using 3x-10x less training cost compared to top-performing Transformer models.
For instance, our model achieves a state-of-the-art result on theenwik8dataset using 1.6 days of training on an 8-GPU machine.
We further demonstrate that SRU++ requires minimal attention for near state-of-the-art performance.
Our results suggest jointly leveraging fast recurrence with little attention as a promising direction for accelerating model training and inference.111Our code, experimental setup and models are available athttps://github.com/asappresearch/sru.
|
2205.11588
|
In this work, we explore whether modeling recurrence into the Transformer architecture can both be beneficial and efficient,
by building an extremely simple recurrent module into the Transformer.
We compare our model to baselines following the training and evaluation recipe of BERT.
Our results confirm that recurrence can indeed improve Transformer models by a consistent margin, without requiring low-level performance optimizations, and while keeping the number of parameters constant.
For example, our base model achieves an absolute improvement of 2.1 points averaged across 10 tasks and also demonstrates increased stability in fine-tuning over a range of learning rates.
|
2312.15166
|
We introduce SOLAR 10.7B, a large language model (LLM) with 10.7 billion parameters, demonstrating superior performance in various natural language processing (NLP) tasks. Inspired by recent efforts to efficiently up-scale LLMs, we present a method for scaling LLMs called depth up-scaling (DUS) , which encompasses depthwise scaling and continued pretraining. In contrast to other LLM up-scaling methods that use mixture-of-experts, DUS does not require complex changes to train and inference efficiently. We show experimentally that DUS is simple yet effective in scaling up high-performance LLMs from small ones. Building on the DUS model, we additionally present SOLAR 10.7B-Instruct, a variant fine-tuned for instruction-following capabilities, surpassing Mixtral-8x7B-Instruct. SOLAR 10.7B is publicly available under the Apache 2.0 license, promoting broad access and application in the LLM field111https://huggingface.co/upstage/SOLAR-10.7B-v1.0.
|
2312.10193
|
The computational cost of transformer models makes them inefficient in low-latency or low-power applications. While techniques such as quantization or linear attention can reduce the computational load, they may incur a reduction in accuracy. In addition, globally reducing the cost for all inputs may be sub-optimal.
We observe that for each layer, the full width of the layer may be needed only for a small subset of tokens inside a batch and that the “effective” width needed to process a token can vary from layer to layer.
Motivated by this observation, we introduce theAdaptive Computation Module(ACM) , a generic module that dynamically adapts its computational load to match the estimated difficulty of the input on a per-token basis.
An ACM consists of a sequence oflearnersthat progressively refine the output of their preceding counterparts. An additional gating mechanism determines the optimal number of learners to execute for each token.
We also describe a distillation technique to replace any pre-trained model with an “ACMized” variant. The distillation phase is designed to be highly parallelizable across layers while being simple to plug-and-play into existing networks.
Our evaluation of transformer models in computer vision and speech recognition demonstrates that substituting layers with ACMs significantly reduces inference costs without degrading the downstream accuracy for a wide interval of user-defined budgets.
|
2311.03687
|
Large Language Models (LLMs) have seen great advance in both academia and industry, and their popularity results in numerous open-source frameworks and techniques in accelerating LLM pre-training, fine-tuning, and inference. Training and deploying LLMs are expensive as it requires considerable computing resources and memory, hence many efficient approaches have been developed for improving system pipelines as well as operators. However, the runtime performance can vary significantly across hardware and software stacks, which makes it difficult to choose the best configuration. In this work, we aim to benchmark the performance from both macro and micro perspectives. First, we benchmark the end-to-end performance of pre-training, fine-tuning, and serving LLMs in different sizes , i.e., 7, 13, and 70 billion parameters (7B, 13B, and 70B) on three 8-GPU platforms with and without individual optimization techniques, including ZeRO, quantization, recomputation, FlashAttention. Then, we dive deeper to provide a detailed runtime analysis of the sub-modules, including computing and communication operators in LLMs. For end users, our benchmark and findings help better understand different optimization techniques, training and inference frameworks, together with hardware platforms in choosing configurations for deploying LLMs. For researchers, our in-depth module-wise analyses discover potential opportunities for future work to further optimize the runtime performance of LLMs.
|
2302.11521
|
Fine-tuning large language models is becoming ever more impractical due to their rapidly-growing scale. This motivates the use of parameter-efficient adaptation methods such as prompt tuning (PT) , which adds a small number of tunable embeddings to an otherwise frozen model, and in-context learning (ICL) , in which demonstrations of the task are provided to the model in natural language without any additional training. Recently,propose “instruction prompt tuning” (IPT) , which combines PT with ICL by concatenating a natural language demonstration with learned prompt embeddings. While all of these methods have proven effective on different tasks, how they interact with each other remains unexplored. In this paper, we empirically study when and how in-context examples improve prompt tuning by measuring the effectiveness of ICL, PT, and IPT on five text generation tasks with multiple base language models. We observe that
(1) IPT doesnotalways outperform PT, and in fact requires the in-context demonstration to be semantically similar to the test input to yield improvements;
(2) PT is unstable and exhibits high variance, but combining PT and ICL (into IPT) consistently reduces variance across all five tasks; and
(3) prompts learned for a specific source task via PT exhibit positive transfer when paired with in-context examples of a different target task.
Our results offer actionable insights on choosing a suitable parameter-efficient adaptation method for a given task.
|
2312.13558
|
Transformer-based Large Language Models (LLMs) have become a fixture in modern machine learning.
Correspondingly, significant resources are allocated towards research that aims to further advance this technology, typically resulting in models of increasing size that are trained on increasing amounts of data.
This work, however, demonstrates the surprising result that it is often possible to significantly improve the performance of LLMs by selectively removing higher-order components111Higher-order components are singular vectors with smaller singular values.of their weight matrices. This simple intervention, which we call LAyer-SElective Rank reduction () , can be done on a model after training has completed, and requires no additional parameters or data. We show extensive experiments demonstrating the generality of this finding across language models and datasets, and provide in-depth analyses offering insights into both whenis effective and the mechanism by which it operates222Code and website:https://pratyushasharma.github.io/laser/.
|
2312.09979
|
Supervised fine-tuning (SFT) is a crucial step for large language models (LLMs) , enabling them to align with human instructions and enhance their capabilities in downstream tasks.
When the models are required to align with a broader range of downstream tasks, or there is a desire to notably improve the performance on a specific task, a substantial increase in fine-tuning data often emerges as the solution.
However, we find that large-scale increases in instruction data can disrupt the world knowledge previously stored in the LLMs, i.e., world knowledge forgetting.
In this paper, we introduce LoRAMoE to address the above challenge.
The LoRAMoE is a plugin version of Mixture of Experts (MoE) .
The plugin form ensures the integrity of world knowledge by freezing the backbone model during the training phase.
We then propose the use of localized balancing constraints to coordinate parts of experts for task utilization, meanwhile enabling other experts to fully leverage the world knowledge stored in the models.
Experimental results demonstrate that LoRAMoE can reasonably coordinate experts based on data type during inference, and even dramatically increasing instruction data does not result in knowledge forgetting.
Moreover, LoRAMoE provides additional benefits for the performance of downstream tasks, indicating the potential of our approach for multi-task learning.
|
2310.00426
|
The most advanced text-to-image (T2I) models require significant training costs (e.g., millions of GPU hours) , seriously hindering the fundamental innovation for the AIGC community while increasing CO2emissions.
This paper introducesPixArt-, a Transformer-based T2I diffusion model whose image generation quality is competitive with state-of-the-art image generators (e.g., Imagen, SDXL, and even Midjourney) , reaching near-commercial application standards. Additionally, it supports high-resolution image synthesis up to 10241024 resolution with low training cost, as shown in Figure1and4.
To achieve this goal, three core designs are proposed:
(1) Training strategy decomposition: We devise three distinct training steps that respectively optimize pixel dependency, text-image alignment, and image aesthetic quality;
(2) Efficient T2I Transformer: We incorporate cross-attention modules into Diffusion Transformer (DiT) to inject text conditions and streamline the computation-intensive class-condition branch;
(3) High-informative data:
We emphasize the significance of concept density in text-image pairs and leverage a large Vision-Language model to auto-label dense pseudo-captions to assist text-image alignment learning.
As a result,PixArt-’s training speed markedly surpasses existing large-scale T2I models,e.g.,PixArt-only takes12%of Stable Diffusion v1.5’s training time (753vs.6,250 A100 GPU days) , saving nearly $300,000 ($28,400vs.$320,000) and reducing 90% CO2emissions.
Moreover, compared with a larger SOTA model, RAPHAEL, our training cost is merely 1%.
Extensive experiments demonstrate thatPixArt-excels in image quality, artistry, and semantic control.
We hopePixArt-will provide new insights to the AIGC community and startups to accelerate building their own high-quality yet low-cost generative models from scratch.
|
2307.14995
|
We present TransNormerLLM, the first linear attention-based Large Language Model (LLM) that outperforms conventional softmax attention-based models in terms of both accuracy and efficiency. TransNormerLLM evolves from the previous linear attention architecture TransNormerby making advanced modifications that include positional embedding, linear attention acceleration, gating mechanism, tensor normalization, and inference acceleration and stabilization.
Specifically, we use LRPEtogether with an exponential decay to avoid attention dilution issues while allowing the model to retain global interactions between tokens.
Additionally, we propose Lightning Attention, a cutting-edge technique that accelerates linear attention by more than twice in runtime and reduces memory usage by a remarkable four times.
To further enhance the performance of TransNormer, we leverage a gating mechanism to smooth training and a new tensor normalization scheme to accelerate the model, resulting in an impressive acceleration of over.
Furthermore, we develop a robust inference algorithm that ensures numerical stability and consistent inference speed, regardless of the sequence length, showcasing superior efficiency during both training and inference stages.
We also implement an efficient model parallel schema for TransNormerLLM, enabling seamless deployment on large-scale clusters and facilitating expansion to even more extensive models,i.e.,LLMs with 175B parameters. We validate our model design through a series of ablations and train models with sizes of 385M, 1B, and 7B on our self-collected corpus. Benchmark results demonstrate that our models not only match the performance of state-of-the-art LLMs with Transformer but are also significantly faster.
|
2312.10431
|
Score-based generative models (or diffusion models for short) have proven successful across many domains in generating text and image data.
However, the consideration of mixed-type tabular data with this model family has fallen short so far.
Existing research mainly combines different diffusion processes without explicitly accounting for the feature heterogeneity inherent to tabular data.
In this paper, we combine score matching and score interpolation to ensure a common type of continuous noise distribution that affects both continuous and categorical features alike.
Further, we investigate the impact of distinct noise schedules per feature or per data type.
We allow for adaptive, learnable noise schedules to ensure optimally allocated model capacity and balanced generative capability.
Results show that our model consistently outperforms state-of-the-art benchmark models and that accounting for heterogeneity within the noise schedule design boosts the sample quality.
|
2312.12682
|
In AI research, the optimization of Large Language Models (LLMs) remains a significant challenge, crucial for advancing the field’s practical applications and sustainability. Building upon the foundational work of Professor Song Han’s lab at MIT, this paper introduces a novel approach in developing Mini-GPTs via contextual pruning. Our methodology strategically prunes the computational architecture of traditional LLMs, like Phi-1.5, focusing on retaining core functionalities while drastically reducing model sizes. We employ the technique across diverse and complex datasets, including US law, Medical Q&A, Skyrim dialogue, English-Taiwanese translation, and Economics articles. The results underscore the efficiency and effectiveness of contextual pruning, not merely as a theoretical concept but as a practical tool in developing domain-specific, resource-efficient LLMs. Contextual pruning is a promising method for building domain-specific LLMs, and this research is a building block towards future development with more hardware compute, refined fine-tuning, and quantization.
|
2312.11514
|
Large language models (LLMs) are central to modern natural language processing, delivering exceptional performance in various tasks. However, their substantial computational and memory requirements present challenges, especially for devices with limited DRAM capacity. This paper tackles the challenge of efficiently running LLMs that exceed the available DRAM capacity by storing the model parameters in flash memory, but bringing them on demand to DRAM.
Our method involves constructing an inference cost model that takes into account the characteristics of flash memory, guiding us to optimize in two critical areas: reducing the volume of data transferred from flash and reading data in larger, more contiguous chunks. Within this hardware-informed framework, we introduce two principal techniques. First, “windowing” strategically reduces data transfer by reusing previously activated neurons, and second, “row-column bundling”, tailored to the sequential data access strengths of flash memory, increases the size of data chunks read from flash memory. These methods collectively enable running models up to twice the size of the available DRAM, with a 4-5x and 20-25x increase in inference speed compared to naive loading approaches in CPU and GPU, respectively. Our integration of sparsity awareness, context-adaptive loading, and a hardware-oriented design paves the way for effective inference of LLMs on devices with limited memory.
|
2312.02696
|
Diffusion models currently dominate the field of datadriven image synthesis with their unparalleled scaling to large datasets. In this paper, we identify and rectify several causes for uneven and ineffective training in the popular ADM diffusion model architecture, without altering its highlevel structure. Observing uncontrolled magnitude changes and imbalances in both the network activations and weights over the course of training, we redesign the network layers to preserve activation, weight, and update magnitudes on expectation. We find that systematic application of this philosophy eliminates the observed drifts and imbalances, resulting in considerably better networks at equal computational complexity. Our modifications improve the previous record FID of 2.41 in ImageNet-512 synthesis to 1.81, achieved using fast deterministic sampling. As an independent contribution, we present a method for setting the exponential moving average (EMA) parameters post-hoc, i.e., after completing the training run. This allows precise tuning of EMA length without the cost of performing several training runs, and reveals its surprising interactions with network architecture, training time, and guidance.
|
2312.09323
|
The goal of this series is to chronicle opinions and issues in the field of machine learning as they stand today and as they change over time. The plan is to host this survey periodically until the AI singularity paperclip-frenzy-driven doomsday, keeping an updated list of topical questions and interviewing new community members for each edition. In this issue, we probed people’s opinions on interpretable AI, the value of benchmarking in modern NLP, the state of progress towards understanding deep learning, and the future of academia.
|
2312.10003
|
Answering complex natural language questions often necessitates multi-step reasoning and integrating external information. Several systems have combined knowledge retrieval with a large language model (LLM) to answer such questions. These systems, however, suffer from various failure cases, and we cannot directly train them end-to-end to fix such failures, as interaction with external knowledge is non-differentiable. To address these deficiencies, we define a ReAct-style LLM agent with the ability to reason and act upon external knowledge. We further refine the agent through a ReST-like method that iteratively trains on previous trajectories, employing growing-batch reinforcement learning with AI feedback for continuous self-improvement and self-distillation. Starting from a prompted large model and after just two iterations of the algorithm, we can produce a fine-tuned small model that achieves comparable performance on challenging compositional question-answering benchmarks with two orders of magnitude fewer parameters.
|
1911.01547
|
To make deliberate progress towards more intelligent and more human-like artificial systems, we need to be following an appropriate feedback signal: we need to be able to define and evaluate intelligence in a way that enables comparisons between two systems, as well as comparisons with humans. Over the past hundred years, there has been an abundance of attempts to define and measure intelligence, across both the fields of psychology and AI. We summarize and critically assess these definitions and evaluation approaches, while making apparent the two historical conceptions of intelligence that have implicitly guided them. We note that in practice, the contemporary AI community still gravitates towards benchmarking intelligence by comparing theskillexhibited by AIs and humans at specific tasks, such as board games and video games. We argue that solely measuring skill at any given task falls short of measuring intelligence, because skill is heavily modulated by prior knowledge and experience: unlimited priors or unlimited training data allow experimenters to “buy” arbitrary levels of skills for a system, in a way that masks the system’s own generalization power. We then articulate a new formal definition of intelligence based on Algorithmic Information Theory, describing intelligence asskill-acquisition efficiencyand highlighting the concepts ofscope,generalization difficulty,priors, andexperience, as critical pieces to be accounted for in characterizing intelligent systems. Using this definition, we propose a set of guidelines for what a general AI benchmark should look like. Finally, we present a new benchmark closely following these guidelines, the Abstraction and Reasoning Corpus (ARC) , built upon an explicit set of priors designed to be as close as possible to innate human priors. We argue that ARC can be used to measure a human-like form of general fluid intelligence and that it enables fair general intelligence comparisons between AI systems and humans.
|
1809.04506
|
In the quest for efficient and robust reinforcement learning methods, both model-free and model-based approaches offer advantages. In this paper we propose a new way of explicitly bridging both approaches via a shared low-dimensional learned encoding of the environment, meant to capture summarizing abstractions. We show that the modularity brought by this approach leads to good generalization while being computationally efficient, with planning happening in a smaller latent state space. In addition, this approach recovers a sufficient low-dimensional representation of the environment, which opens up new strategies for interpretable AI, exploration and transfer learning.
|
2310.06816
|
How much private information do text embeddings reveal about the original text? We investigate the problem of embeddinginversion, reconstructing the full text represented in dense text embeddings. We frame the problem as controlled generation: generating text that, when reembedded, is close to a fixed point in latent space. We find that although a naïve model conditioned on the embedding performs poorly, a multi-step method that iteratively corrects and re-embeds text is able to recoveroftext inputs exactly. We train our model to decode text embeddings from two state-of-the-art embedding models, and also show that our model can recover important personal information (full names) from a dataset of clinical notes.111Our code is available on Github:github.com/jxmorris12/vec2text.
|
2305.19118
|
Modern large language models (LLMs) like ChatGPT have shown remarkable performance on general language tasks but still struggle on complex reasoning tasks, which drives the research on cognitive behaviors of LLMs to explore human-like problem-solving strategies.
Along this direction, one representative strategy is self-reflection, which asks an LLM to refine the solution with the feedback generated by itself iteratively.
However, our study shows that such reflection-style methods suffer from theDegeneration-of-Thought (DoT) problem: once the LLM has established confidence in its solutions, it is unable to generate novel thoughts later through reflection even if its initial stance is incorrect.
To address the DoT problem, we propose aMulti-Agent Debate (MAD) framework, in which multiple agents express their arguments in the state of “tit for tat” and a judge manages the debate process to obtain a final solution.
Clearly, our MAD framework encourages divergent thinking in LLMs which would be helpful for tasks that require deep levels of contemplation.
Experiment results on two challenging datasets, commonsense machine translation and counter-intuitive arithmetic reasoning, demonstrate the effectiveness of our MAD framework.
Extensive analyses suggest that the adaptive break of debate and the modest level of “tit for tat” state are required for MAD to obtain good performance. Moreover, we find that LLMs might not be a fair judge if different LLMs are used for agents. Codes:https://github.com/Skytliang/Multi-Agents-Debate
|
1907.05242
|
This paper introduces a structured memory which can be easily integrated into a neural network.
The memory is very large by design and significantly increases the capacity of the architecture, by up to a billion parameters with a negligible computational overhead.
Its design and access pattern is based on product keys, which enable fast and exact nearest neighbor search.
The ability to increase the number of parameters while keeping the same computational budget lets the overall system strike a better trade-off between prediction accuracy and computation efficiency both at training and test time.
This memory layer allows us to tackle very large scale language modeling tasks. In our experiments we consider a dataset with up to 30 billion words, and we plug our memory layer in a state-of-the-art transformer-based architecture.
In particular, we found that a memory augmented model with only 12 layers outperforms a baseline transformer model with 24 layers, while being twice faster at inference time. We release our code for reproducibility purposes.111https://github.com/facebookresearch/XLM
|
2303.15647
|
This paper presents a systematic overview and comparison of parameter-efficient fine-tuning methods covering over 40 papers published between February 2019 and February 2023. These methods aim to resolve the infeasibility and impracticality of fine-tuning large language models by only training a small set of parameters. We provide a taxonomy that covers a broad range of methods and present a detailed method comparison with a specific focus on real-life efficiency and fine-tuning multibillion-scale language models.
|
2312.09241
|
Small-scale models offer various computational advantages, and yet to which extent size is critical for problem-solving abilities remains an open question. Specifically for solving grade school math, the smallest model size so far required to break the 80% barrier on the GSM8K benchmark remains to be 34B.
Our work studies how high-quality datasets may be the key for small language models to acquire mathematical reasoning. We introduceTinyGSM, a synthetic dataset of 12.3M grade school math problems paired with Python solutions, generated fully by GPT-3.5. After finetuning onTinyGSM, we find that a duo of a 1.3B generation model and a 1.3B verifier model can achieve 81.5% accuracy, outperforming existing models that are orders of magnitude larger.
This also rivals the performance of the GPT-3.5 “teacher” model (77.4%) , from which our model’s training data is generated.
Our approach is simple and has two key components:
1) the high-quality datasetTinyGSM,
2) the use of a verifier, which selects the final outputs from multiple candidate generations.
|
2212.05055
|
Training large, deep neural networks to convergence can be prohibitively expensive. As a result, often only a small selection of popular, dense models are reused across different contexts and tasks.
Increasingly, sparsely activated models, which seek to decouple model size from computation costs, are becoming an attractive alternative to dense models. Although more efficient in terms of quality and computation cost, sparse models remain data-hungry and costly to train from scratch in the large scale regime.
In this work, we propose sparse upcycling – a simple way to reuse sunk training costs by initializing a sparsely activated Mixture-of-Experts model from a dense checkpoint.
We show that sparsely upcycled T5 Base, Large, and XL language models and Vision Transformer Base and Large models, respectively, significantly outperform their dense counterparts on SuperGLUE and ImageNet, using onlyof the initial dense pretraining sunk cost. The upcycled models also outperform sparse models trained from scratch onof the initial dense pretraining computation budget.111Code is available athttps://github.com/google-research/vmoe(Vision) andhttps://github.com/google-research/t5x/tree/main/t5x/contrib/moe(Language) .
|
2103.10334
|
Language model pre-training and derived methods are incredibly impactful in machine learning.
However, there remains considerable uncertainty on exactly why pre-training helps improve performance for fine-tuning tasks. This is especially true when attempting to adapt language-model pre-training to domains outside of natural language.
Here, we analyze this problem by exploring how existing pre-training methods impose relational structure in their induced per-sample latent spaces—i.e., what constraints do pre-training methods impose on the distance or geometry between the pre-trained embeddings of two samplesand.
Through a comprehensive review of existing pre-training methods, we find that this question remains open. This is true despite theoretical analyses demonstrating the importance of understanding this form of induced structure.
Based on this review, we introduce a descriptive framework for pre-training that allows for a granular, comprehensive understanding of how relational structure can be induced. We present a theoretical analysis of this framework from first principles and establish a connection between the relational inductive bias of pre-training and fine-tuning performance. We also show how to use the framework to define new pre-training methods.
We build upon these findings with empirical studies on benchmarks spanning 3 data modalities and ten fine-tuning tasks. These experiments validate our theoretical analyses, inform the design of novel pre-training methods, and establish consistent improvements over a compelling suite of baseline methods.
|
2309.16298
|
Large Language Models (LLMs) have exhibited remarkable reasoning capabilities and become the foundation of language technologies. Inspired by the great success of code data in training LLMs, we naturally wonder at which training stage introducing code data can really help LLMs reasoning. To this end, this paper systematically explores the impact of code data on LLMs at different stages. Concretely, we introduce the code data at the pre-training stage, instruction-tuning stage, and both of them, respectively. Then, the reasoning capability of LLMs is comprehensively and fairly evaluated via six reasoning tasks in five domains. We critically analyze the experimental results and provide conclusions with insights. First, pre-training LLMs with the mixture of code and text can significantly enhance LLMs’ general reasoning capability almost without negative transfer on other tasks. Besides, at the instruction-tuning stage, code data endows LLMs the task-specific reasoning capability. Moreover, the dynamic mixing strategy of code and text data assists LLMs to learn reasoning capability step-by-step during training. These insights deepen the understanding of LLMs regarding reasoning ability for their application, such as scientific question answering, legal support, etc. The source code and model parameters are released at the link:https://github.com/yingweima2022/CodeLLM.
|
2310.09336
|
Modern generative models exhibit unprecedented capabilities to generate extremely realistic data. However, given the inherent compositionality of the real world, reliable use of these models in practical applications requires that they exhibit the capability to compose a novel set of concepts to generate outputs not seen in the training data set. Prior work demonstrates that recent diffusion models do exhibit intriguing compositional generalization abilities, but also fail unpredictably. Motivated by this, we perform a controlled study for understanding compositional generalization in conditional diffusion models in a synthetic setting, varying different attributes of the training data and measuring the model’s ability to generate samples out-of-distribution. Our results show: (i) the order in which the ability to generate samples from a concept and compose them emerges is governed by the structure of the underlying data-generating process; (ii) performance on compositional tasks exhibits a sudden “emergence” due to multiplicative reliance on the performance of constituent tasks, partially explaining emergent phenomena seen in generative models; and (iii) composing concepts with lower frequency in the training data to generate out-of-distribution samples requires considerably more optimization steps compared to generating in-distribution samples. Overall, our study lays a foundation for understanding emergent capabilities and compositionality in generative models from a data-centric perspective. Our code is available athttps://github.com/phys-ai/concept_graphs.
|
2312.00752
|
Foundation models, now powering most of the exciting applications in deep learning, are almost universally based on the Transformer architecture and its core attention module. Many subquadratic-time architectures such as linear attention, gated convolution and recurrent models, and structured state space models (SSMs) have been developed to address Transformers' computational inefficiency on long sequences, but they have not performed as well as attention on important modalities such as language. We identify that a key weakness of such models is their inability to perform content-based reasoning, and make several improvements. First, simply letting the SSM parameters be functions of the input addresses their weakness with discrete modalities, allowing the model to *selectively* propagate or forget information along the sequence length dimension depending on the current token. Second, even though this change prevents the use of efficient convolutions, we design a hardware-aware parallel algorithm in recurrent mode. We integrate these selective SSMs into a simplified end-to-end neural network architecture without attention or even MLP blocks (**Mamba**). Mamba enjoys fast inference (5× higher throughput than Transformers) and linear scaling in sequence length, and its performance improves on real data up to million-length sequences. As a general sequence model backbone, Mamba achieves state-of-the-art performance across several modalities such as language, audio, and genomics. On language modeling, our Mamba-3B model outperforms Transformers of the same size and matches Transformers twice its size, both in pretraining and downstream evaluation.
|
2309.14327
|
Most of the existing multi-modal models, hindered by their incapacity to adeptly manage interleaved image-and-text inputs in multi-image, multi-round dialogues, face substantial constraints in resource allocation for training and data accessibility, impacting their adaptability and scalability across varied interaction realms.
To address this, we present the DeepSpeed-VisualChat framework, designed to optimize Large Language Models (LLMs) by incorporating multi-modal capabilities, with a focus on enhancing the proficiency of Large Vision and Language Models in handling interleaved inputs.
Our framework is notable for (1) its open-source support for multi-round and multi-image dialogues, (2) introducing an innovative multi-modal casual attention mechanism, and (3) utilizing data blending techniques on existing datasets to assure seamless interactions in multi-round, multi-image conversations.
Compared to existing frameworks, DeepSpeed-VisualChat shows superior scalability up to 70B parameter language model size, representing a significant advancement in multi-modal language models and setting a solid foundation for future explorations.111Code will be released soon as a part ofhttps://github.com/microsoft/DeepSpeedExample
|
2311.17541
|
Large Language Models (LLMs) have shown impressive abilities in natural language understanding and generation, leading to their use in applications such as chatbots and virtual assistants. However, existing LLM frameworks face limitations in handling domain-specific data analytics tasks with rich data structures. Moreover, they struggle with flexibility to meet diverse user requirements.
To address these issues, TaskWeaver is proposed as a code-first framework for building LLM-powered autonomous agents. It converts user requests into executable code and treats user-defined plugins as callable functions. TaskWeaver provides support for rich data structures, flexible plugin usage, and dynamic plugin selection, and leverages LLM coding capabilities for complex logic. It also incorporates domain-specific knowledge through examples and ensures the secure execution of generated code. TaskWeaver offers a powerful and flexible framework for creating intelligent conversational agents that can handle complex tasks and adapt to domain-specific scenarios. The code is open-sourced athttps://github.com/microsoft/TaskWeaver/.
|
2311.18823
|
Weight initialization plays an important role in neural network training. Widely used initialization methods are proposed and evaluated for networks that are trained from scratch. However, the growing number of pretrained models now offers new opportunities for tackling this classical problem of weight initialization. In this work, we introduce weight selection, a method for initializing smaller models by selecting a subset of weights from a pretrained larger model. This enables the transfer of knowledge from pretrained weights to smaller models. Our experiments demonstrate that weight selection can significantly enhance the performance of small models and reduce their training time. Notably, it can also be used together with knowledge distillation. Weight selection offers a new approach to leverage the power of pretrained models in resource-constrained settings, and we hope it can be a useful tool for training small models in the large-model era. Code is available athttps://github.com/OscarXZQ/weight-selection.
|
2211.15089
|
Diffusion models have quickly become the go-to paradigm for generative modelling of perceptual signals (such as images and sound) through iterative refinement. Their success hinges on the fact that the underlying physical phenomena are continuous. For inherently discrete and categorical data such as language, various diffusion-inspired alternatives have been proposed. However, the continuous nature of diffusion models conveys many benefits, and in this work we endeavour to preserve it. We propose CDCD, a framework for modelling categorical data with diffusion models that are continuous both in time and input space. We demonstrate its efficacy on several language modelling tasks.
|
2311.18775
|
We present CoDi-2, a versatile and interactive Multimodal Large Language Model (MLLM) that can follow complex multimodal interleaved instructions, conduct in-context learning (ICL) , reason, chat, edit, etc., in an any-to-any input-output modality paradigm. By aligning modalities with language for both encoding and generation, CoDi-2 empowers Large Language Models (LLMs) to not only understand complex modality-interleaved instructions and in-context examples, but also autoregressively generate grounded and coherent multimodal outputs in the continuous feature space.
To train CoDi-2, we build a large-scale generation dataset encompassing in-context multimodal instructions across text, vision, and audio. CoDi-2 demonstrates a wide range of zero-shot capabilities for multimodal generation, such as in-context learning, reasoning, and compositionality of any-to-any modality generation through multi-round interactive conversation. CoDi-2 surpasses
previous domain-specific models on tasks such as subject-driven image generation, vision transformation, and audio editing. CoDi-2 signifies a substantial breakthrough in developing a comprehensive multimodal foundation model adept at interpreting in-context language-vision-audio interleaved instructions and producing multimodal outputs.
|
2311.17311
|
Self-consistency with chain-of-thought prompting (CoT) has demonstrated remarkable performance gains on various challenging tasks, by utilizing multiple reasoning paths sampled from large language models (LLMs) . However, self-consistency relies on the answer extraction process to aggregate multiple solutions, which is not applicable to free-form answers. In this work, we propose Universal Self-Consistency (USC) , which leverages LLMs themselves to select the most consistent answer among multiple candidates. We evaluate USC on a variety of benchmarks, including mathematical reasoning, code generation, long-context summarization, and open-ended question answering. On open-ended generation tasks where the original self-consistency method is not applicable, USC effectively utilizes multiple samples and improves the performance. For mathematical reasoning, USC matches the standard self-consistency performance without requiring the answer formats to be similar. Finally, without access to execution results, USC also matches the execution-based voting performance on code generation.
|
2310.10944
|
As large language models (LLMs) become more prevalent, there is a growing need for new and improved quantization methods that can meet the computationalast layer demands of these modern architectures while maintaining the accuracy. In this paper, we present TEQ, a trainable equivalent transformation that preserves the FP32 precision of the model output while taking advantage of low-precision quantization, especially 3 and 4 bits weight-only quantization. The training process is lightweight, requiring only 1K steps and less thanof the original model’s trainable parameters. Furthermore, the transformation does not add any computational overhead during inference. Our results are on-par with the state-of-the-art (SOTA) methods on typical LLMs. Our approach can be combined with other methods to achieve even better performance. The code is available at https://github.com/intel/neural-compressor.
|
2311.14097
|
Though diffusion models excel in image generation, their step-by-step denoising leads to slow generation speeds. Consistency training addresses this issue with single-step sampling but often produces lower-quality generations and requires high training costs. In this paper, we show that optimizing consistency training loss minimizes the Wasserstein distance between target and generated distributions. As timestep increases, the upper bound accumulates previous consistency training losses. Therefore, larger batch sizes are needed to reduce both current and accumulated losses. We propose Adversarial Consistency Training (ACT) , which directly minimizes the Jensen-Shannon (JS) divergence between distributions at each timestep using a discriminator. Theoretically, ACT enhances generation quality, and convergence. By incorporating a discriminator into the consistency training framework, our method achieves improved FID scores on CIFAR10, ImageNet 6464 and LSUN Cat 256256 datasets, retains zero-shot image inpainting capabilities, and uses less thanof the original batch size and fewer thanof the model parameters and training steps compared to the baseline method, this leads to a substantial reduction in resource consumption.
|
2311.05640
|
Large language models (LLMs) excel in many tasks in NLP and beyond, but most open models have very limited coverage of smaller languages and LLM work tends to focus on languages where nearly unlimited data is available for pretraining. In this work, we study the challenges of creating LLMs for Finnish, a language spoken by less than 0.1% of the world population. We compile an extensive dataset of Finnish combining web crawls, news, social media and eBooks. We pursue two approaches to pretrain models: 1) we train seven monolingual models from scratch (186M to 13B parameters) dubbed FinGPT, 2) we continue the pretraining of the multilingual BLOOM model on a mix of its original training data and Finnish, resulting in a 176 billion parameter model we call BLUUMI.
For model evaluation, we introduce FIN-bench, a version of BIG-bench with Finnish tasks. We also assess other model qualities such as toxicity and bias. Our models and tools are openly available athttps://turkunlp.org/gpt3-finnish.
|
2311.14737
|
Transformers, central to the successes in modern Natural Language Processing, often falter on arithmetic tasks despite their vast capabilities –which paradoxically include remarkable coding abilities. We observe that a crucial challenge is their naive reliance on positional information to solve arithmetic problems with a small number of digits, leading to poor performance on larger numbers. Herein, we delve deeper into the role of positional encoding, and propose several ways to fix the issue, either by modifying the positional encoding directly, or by modifying the representation of the arithmetic task to leverage standard positional encoding differently. We investigate the value of these modifications for three tasks: (i) classical multiplication, (ii) length extrapolation in addition, and (iii) addition in natural language context. For (i) we train a small model on a small dataset (100M parameters and 300k samples) with remarkable aptitude in (direct, no scratchpad) 15 digits multiplication and essentially perfect up to 12 digits, while usual training in this context would give a model failing at 4 digits multiplication. In the experiments on addition, we use a mere 120k samples to demonstrate: for (ii) extrapolation from 10 digits to testing on 12 digits numbers while usual training would have no extrapolation, and for (iii) almost perfect accuracy up to 5 digits while usual training would be correct only up to 3 digits (which is essentially memorization with a training set of 120k samples) .
|
2306.02561
|
We presentLLM-Blender, an ensembling framework designed to attain consistently superior performance by leveraging the diverse strengths of multiple open-source large language models (LLMs) . Our framework consists of two modules:PairRankerandGenFuser, addressing the observation that optimal LLMs for different examples can significantly vary.PairRankeremploys a specialized pairwise comparison method to distinguish subtle differences between candidate outputs. It jointly encodes the input text and a pair of candidates, using cross-attention encoders to determine the superior one. Our results demonstrate thatPairRankerexhibits the highest correlation with ChatGPT-based ranking.
Then,GenFuseraims to merge the top-ranked candidates, generating an improved output by capitalizing on their strengths and mitigating their weaknesses. To facilitate large-scale evaluation, we introduce a benchmark dataset,MixInstruct, which is a mixture of multiple instruction datasets featuring oracle pairwise comparisons.
OurLLM-Blendersignificantly outperform individual LLMs and baseline methods across various metrics, establishing a substantial performance gap.111https://yuchenlin.xyz/LLM-Blender222The experiments on summarization, translation, and constrained generation tasks in the prior version have been moved to the appendix. Instead, we mainly present our work in the context of instruction-following data and LLMs in this version.
|
2311.08401
|
The fluency and creativity of large pre-trained language models (LLMs) have led to their widespread use, sometimes even as a replacement for traditional search engines. Yet language models are prone to making convincing but factually inaccurate claims, often referred to as ‘hallucinations.’ These errors can inadvertently spread misinformation or harmfully perpetuate misconceptions.
Further, manual fact-checking of model responses is a time-consuming process, making human factuality labels expensive to acquire.
In this work, we fine-tune language models to be more factual, without human labeling and targeting more open-ended generation settings than past work. We leverage two key recent innovations in NLP to do so. First, several recent works have proposed methods for judging the factuality of open-ended text by measuring consistency with an external knowledge base or simply a large model’s confidence scores. Second, the direct preference optimization algorithm enables straightforward fine-tuning of language models on objectives other than supervised imitation, using a preference ranking over possible model responses. We show that learning from automatically generated factuality preference rankings, generated either through existing retrieval systems or our novel retrieval-free approach, significantly improves the factuality (percent of generated claims that are correct) of Llama-2 on held-out topics compared with RLHF or decoding strategies targeted at factuality. At 7B scale,compared to Llama-2-chat, we observe 58% and 40% reduction in factual error ratewhen generating biographies and answering medical questions, respectively.
|
2310.12036
|
The prevalent deployment of learning from human preferences through reinforcement learning (RLHF) relies
on two important approximations: the first assumes that pairwise preferences can be substituted with pointwise rewards. The second assumes that a reward model trained on these pointwise rewards can generalize from collected data to out-of-distribution data sampled by the policy. Recently, Direct Preference Optimisation (DPO) has been proposed as an approach that bypasses the second approximation and learn directly a policy from collected data without the reward modelling stage. However, this method still heavily relies on the first approximation. In this paper we try to gain a deeper theoretical understanding of these practical algorithms. In particular we derive a new general objective
calledPOfor learning from human preferences that is expressed in terms of pairwise preferences and therefore bypasses both approximations. This new general objective allows us to perform an in-depth analysis of the behavior ofRLHFandDPO(as special cases ofPO) and to identify their potential pitfalls. We then consider another special case forPOby settingsimply to Identity, for which we can derive an efficient optimisation procedure, prove performance guarantees and demonstrate its
empirical superiority toDPOon some illustrative examples.
|
2311.13171v1
|
Parameter-efficient fine-tuning (PEFT) techniques make it possible to efficiently adapt a language model to create “expert” models that specialize to new tasks or domains. Recent techniques in model merging and compositional generalization leverage these expert models by dynamically composing modules to improve zero/few-shot generalization. Despite the efficiency of PEFT methods, the size of expert models can make it onerous to retrieve expert models per query over high-latency networks like the Internet or serve multiple experts on a single GPU. To address these issues, we present, a novel method for compressing fine-tuning residuals (task vectors) of PEFT-based models.employs sparsification and ternary quantization to reduce the size of the PEFT module without performing any additional retraining while preserving or enhancing model performance. In extensive evaluation across,, and-based models withparameters,achieves compression ratios of. In particular, we show thatimproves with scale – stronger models exhibit higher compressibility and better performance. For example, we showapplied to LLaMA outperforms QLoRA byon MMLU with a storage size reduction of up to. In addition, we show that the compressed experts produced bymaintain few-shot compositional generalization capabilities, facilitate efficient communication and computation, and exhibit enhanced performance when merged. Lastly, we provide an analysis of different method components, compare it with other PEFT methods, and test’s efficacy for compressing the residual of full-finetuning.111Code is available athttps://github.com/prateeky2806/compeft.
|
2206.08954v2
|
Self-supervised learning (SSL) has recently achieved tremendous empirical advancements in learning image representation. However, our understanding of the principle behind learning such a representation is still limited. This work shows that joint-embedding SSL approaches primarily learn a representation of image patches, which reflects their co-occurrence. Such a connection to co-occurrence modeling can be established formally, and it supplements the prevailing invariance perspective. We empirically show that learning a representation for fixed-scale patches and aggregating local patch representations as the image representation achieves similar or even better results than the baseline methods. We denote this process asBagSSL. Even withpatch representation, BagSSL achievestop-1 linear probing accuracy on ImageNet. On the other hand, with a multi-scale pretrained model, we show that the whole image embedding is approximately the average of local patch embeddings. While the SSL representation is relatively invariant at the global scale, we show that locality is preserved when we zoom into local patch-level representation. Further, we show that patch representation aggregation can improve various SOTA baseline methods by a large margin. The patch representation is considerably easier to understand, and this work makes a step to demystify self-supervised representation learning.
|
2311.12023v1
|
We propose a simple approach for memory-efficient adaptation of pretrained language models. Our approach uses an iterative algorithm to decompose each pretrained matrix into a high-precision low-rank component and a memory-efficient quantized component. During finetuning, the quantized component remains fixed and only the low-rank component is updated. We present an integer linear programming formulation of the quantization component which enables dynamic configuration of quantization parameters (e.g., bit-width, block size) for each matrix given an overall target memory budget. We further explore a data-aware version of the algorithm which uses an approximation of the Fisher information matrix to weight the reconstruction objective during matrix decomposition. Experiments on finetuning RoBERTa and LLaMA-2 (7B and 70B) demonstrate that our low-rank plus quantized matrix decomposition approach (LQ-LoRA) outperforms strong QLoRA and GPTQ-LoRA baselines and enables aggressive quantization to sub-3 bits with only minor performance degradations. When finetuned on a language modeling calibration dataset, LQ-LoRA can also be used for model compression; in this setting our 2.75-bit LLaMA-2-70B model (which has 2.85 bits on average when including the low-rank components and requires 27GB of GPU memory) performs respectably compared to the 16-bit baseline.111Our code and models are available athttps://github.com/HanGuo97/lq-lora. This work was completed while Han Guo was a visiting student at MIT.
|
2311.08516v1
|
While self-correction has shown promise in improving LLM outputs in terms of style and quality, recent attempts to self-correct logical or reasoning errors often cause correct answers to become incorrect, resulting in worse performances overall.
In this paper, we break down the self-correction process into two core components:mistake findingandoutput correction. For mistake finding, we release BIG-Bench Mistake, a dataset of logical mistakes in Chain-of-Thought reasoning traces. We provide benchmark numbers for several state-of-the-art LLMs, and demonstrate that LLMs generally struggle with finding logical mistakes. For output correction, we propose a backtracking method which provides large improvements when given information on mistake location. We construe backtracking as a lightweight alternative to reinforcement learning methods, and show that it remains effective with a reward model at 60-70% accuracy.
|
2311.09578v1
|
We propose Tied-LoRA , a simple paradigm utilizes weight tying and selective training to further increase parameter efficiency of the Low-rank adaptation (LoRA) method.
Our investigations include all feasible combinations parameter training/freezing in conjunction with weight tying to identify the optimal balance between performance and the number of trainable parameters. Through experiments covering a variety of tasks and two base language models, we provide analysis revealing trade-offs between efficiency and performance. Our experiments uncovered a particular Tied-LoRA configuration that stands out by demonstrating comparable performance across several tasks while employing only 13 % percent of parameters utilized by the standard LoRA method.
|
2311.10708v1
|
In this work, we show that text-to-image generative models can be ‘inverted’ to assess their own text-image understanding capabilities in a completely automated manner.
Our method, calledSelfEval, uses the generative model to compute the likelihood of real images given text prompts, making the generative model directly applicable to discriminative tasks.
UsingSelfEval, we repurpose standard datasets created for evaluating multimodal text-image discriminative models to evaluate generative models in a fine-grained manner: assessing their performance on attribute binding, color recognition, counting, shape recognition, spatial understanding.
To the best of our knowledgeSelfEvalis the first automated metric to show a high degree of agreement for measuring text-faithfulness with the gold-standard human evaluations across multiple models and benchmarks.
Moreover,SelfEvalenables us to evaluate generative models on challenging tasks such as Winoground image-score where they demonstrate competitive performance to discriminative models.
We also show severe drawbacks of standard automated metrics such as CLIP-score to measure text faithfulness on benchmarks such as DrawBench, and howSelfEvalsidesteps these issues.
We hopeSelfEvalenables easy and reliable automated evaluation for diffusion models.
|
2311.09247v1
|
We explore the abstract reasoning abilities of text-only and multimodal versions of GPT-4, using the ConceptARC benchmark, which is designed to evaluate robust understanding and reasoning with core-knowledge concepts. We extend the work of Moskvichev et al.by evaluating GPT-4 on more detailed, one-shot prompting (rather than simple, zero-shot prompts) with text versions of ConceptARC tasks, and by evaluating GPT-4V, the multimodal version of GPT-4, on zero- and one-shot prompts using image versions of the simplest tasks. Our experimental results support the conclusion that neither version of GPT-4 has developed robust abstraction abilities at humanlike levels.
|
2311.00176v2
|
ChipNeMo aims to explore the applications of large language models (LLMs) for industrial chip design. Instead of directly deploying off-the-shelf commercial or open-source LLMs, we instead adopt the following domain adaptation techniques: custom tokenizers, domain-adaptive continued pretraining, supervised fine-tuning (SFT) with domain-specific instructions, and domain-adapted retrieval models. We evaluate these methods on three selected LLM applications for chip design: an engineering assistant chatbot, EDA script generation, and bug summarization and analysis.
Our results show that these domain adaptation techniques enable significant LLM performance improvements over general-purpose base models across the three evaluated applications, enabling up to 5x model size reduction with similar or better performance on a range of design tasks.
Our findings also indicate that there’s still room for improvement between our current results and ideal outcomes. We believe that further investigation of domain-adapted LLM approaches will help close this gap in the future.
|
2311.04163
|
We identify a new phenomenon in neural network optimization which arises from the interaction of depth and a particular heavy-tailed structure in natural data. Our result offers intuitive explanations for several previously reported observations about network training dynamics. In particular, it implies a conceptually new cause for progressive sharpening and the edge of stability; we also highlight connections to other concepts in optimization and generalization including grokking, simplicity bias, and Sharpness-Aware Minimization. Experimentally, we demonstrate the significant influence of paired groups of outliers in the training data with strongopposing signals: consistent, large magnitude features which dominate the network output throughout training and provide gradients which point in opposite directions. Due to these outliers, early optimization enters a narrow valley which carefully balances the opposing groups; subsequent sharpening causes their loss to rise rapidly, oscillating between high on one group and then the other, until the overall loss spikes. We describe how to identify these groups, explore what sets them apart, and carefully study their effect on the network’s optimization and behavior.
We complement these experiments with a mechanistic explanation on a toy example of opposing signals and a theoretical analysis of a two-layer linear network on a simple model.
Our finding enables new qualitative predictions of training behavior which we confirm experimentally. It also provides a new lens through which to study and improve modern training practices for stochastic optimization, which we highlight via a case study of Adam versus SGD.
|
2305.20050
|
In recent years, large language models have greatly improved in their ability to perform complex multi-step reasoning. However, even state-of-the-art models still regularly produce logical mistakes. To train more reliable models, we can turn either to outcome supervision, which provides feedback for a final result, or process supervision, which provides feedback for each intermediate reasoning step. Given the importance of training reliable models, and given the high cost of human feedback, it is important to carefully compare the both methods. Recent work has already begun this comparison, but many questions still remain. We conduct our own investigation, finding that process supervision significantly outperforms outcome supervision for training models to solve problems from the challenging MATH dataset. Our process-supervised model solves 78% of problems from a representative subset of the MATH test set. Additionally, we show that active learning significantly improves the efficacy of process supervision. To support related research, we also release PRM800K, the complete dataset of 800,000 step-level human feedback labels used to train our best reward model.
|
2311.10207
|
From classical HPC to deep learning,Matrix Multiplication (MatMul) is at the heart of today’s computing.
The recentMaddness111Multiply-ADDitioN-lESSmethod approximatesMatMulwithout the need for multiplication by using a hash-based version of product quantization (PQ) indexing into a look-up table (Look-Up Table (LUT) ) .Stella Nera222Named after an album of the famous Swiss musicians Patent Ochsneris the firstMaddnessaccelerator and it achieves 15x higher area efficiency (GMAC/s/mm2) and a more than 25x higher energy efficiency (TMAC/s/W) than direct MatMul accelerators implemented in the same technology. The hash function is a decision tree, which allows for an efficient hardware implementation as the multiply-accumulate operations are replaced by decision tree passes andLUTlookups. The entireMaddnessMatMul can be broken down into parts that allow an effective implementation with small computing units and memories, allowing it to reach extreme efficiency while remaining generically applicable for MatMul tasks.
In a commercial 14 nm technology and scaled to 3 nm, we achieve an energy efficiency of 161 TOp/s/[email protected] with a Top-1 accuracy on CIFAR-10 of more than 92.5% using ResNet9.
|
2311.08105
|
Large language models (LLM) have become a critical component in many applications of machine learning. However, standard approaches to training LLM require a large number of tightly interconnected accelerators, with devices exchanging gradients and other intermediate states at each optimization step. While it is difficult to build and maintain a single computing cluster hosting many accelerators, it might be easier to find several computing clusters each hosting a smaller number of devices. In this work, we propose a distributed optimization algorithm, Distributed Low-Communication (DiLoCo) , that enables training of language models on islands of devices that are poorly connected. The approach is a variant of federated averaging, where the number of inner steps is large, the inner optimizer is AdamW, and the outer optimizer is Nesterov momentum. On the widely used C4 dataset, we show that DiLoCo on 8 workers performs as well as fully synchronous optimization while communicating 500 times less. DiLoCo exhibits great robustness to the data distribution of each worker. It is also robust to resources becoming unavailable over time, and vice versa, it can seamlessly leverage resources that become available during training.
|
2306.04751
|
In this work we explore recent advances in instruction-tuning language models on a range of open instruction-following datasets.
Despite recent claims that open models can be on par with state-of-the-art proprietary models,
these claims are often accompanied by limited evaluation, making it difficult to compare models across the board and determine the utility of various
resources. We provide a large set of instruction-tuned models
from 6.7B to 65B parameters in size, trained on 12 instruction datasets ranging from manually curated (e.g., OpenAssistant) to synthetic and distilled (e.g., Alpaca) and systematically evaluate them on their factual knowledge, reasoning, multilinguality, coding, safety, and open-ended instruction following abilities through a collection of automatic, model-based, and human-based metrics.
We further introduceTülu, our best performing instruction-tuned model suite finetuned on a combination of high-quality open resources. Our experiments show that different instruction-tuning datasets can uncover or enhance specific skills, while no single dataset (or combination) provides the best performance across all evaluations.
Interestingly, we find that model and human preference-based evaluations fail to reflect differences in model capabilities exposed by benchmark-based evaluations, suggesting the need for the type of systemic evaluation performed in this work.
Our evaluations show that the best model in any given evaluation reaches on average 87% of ChatGPT performance, and 73% of GPT-4 performance, suggesting that further investment in building better base models and instruction-tuning data is required to close the gap.
We release our instruction-tuned models, including a fully finetuned 65BTülu,
along with our code, data, and evaluation framework to facilitate future research.111https://github.com/allenai/open-instruct
|
2311.10702
|
Since the release ofTülu, open resources for instruction tuning have developed quickly, from better base models to new finetuning techniques. We test and incorporate a number of these advances intoTülu, resulting inTülu2, a suite of improvedTülumodels for advancing the understanding and best practices of adapting pretrained language models to downstream tasks and user preferences. Concretely, we release: (1) Tülu-V2-mix, an improved collection of high-quality instruction datasets; (2) Tülu2,Llama-2 models finetuned on the V2 mixture; (3) Tülu2+DPO,Tülu2 models trained with direct preference optimization (DPO) , including the largest DPO-trained model to date (Tülu2+DPO 70B) ; (4) Code Tülu2,Code Llamamodels finetuned on our V2 mix that outperformCode Llamaand its instruction-tuned variant,Code Llama-Instruct. Our evaluation from multiple perspectives shows that theTülu2 suite achieves state-of-the-art performance among open models and matches or exceeds the performance of GPT-3.5-turbo-0301 on several benchmarks. We release all the checkpoints, data, training and evaluation code to facilitate future open efforts on adapting large language models.
|
2311.11077
|
We introduceAdapters, an open-source library that unifies parameter-efficient and modular transfer learning in large language models. By integrating 10 diverse adapter methods into a unified interface,Adaptersoffers ease of use and flexible configuration. Our library allows researchers and practitioners to leverage adapter modularity through composition blocks, enabling the design of complex adapter setups. We demonstrate the library’s efficacy by evaluating its performance against full fine-tuning on various NLP tasks.Adaptersprovides a powerful tool for addressing the challenges of conventional fine-tuning paradigms and promoting more efficient and modular transfer learning. The library is available viahttps://adapterhub.ml/adapters.
|
2311.11829
|
Soft attention in Transformer-based Large Language Models (LLMs) is
susceptible to incorporating irrelevant information from the context into its latent representations, which adversely affects next token generations.
To help rectify these issues, we introduce System 2 Attention (S2A) , which leverages the ability of LLMs to reason in natural language and follow instructions
in order to decide what to attend to. S2A regenerates the input context to only include the relevant portions, before attending to the regenerated context to elicit the final response.
In experiments, S2A outperforms standard attention-based LLMs on
three tasks containing opinion or irrelevant information: QA, math word problems and longform generation,
where S2A increases factuality and objectivity, and decreases sycophancy.
|
2311.07911
|
One core capability of Large Language Models (LLMs) is to follow natural language instructions. However, the evaluation of such abilities is not standardized: Human evaluations are expensive, slow, and not objectively reproducible, while LLM-based auto-evaluation is potentially biased or limited by the ability of the evaluator LLM. To overcome these issues, we introduce Instruction-Following Eval (IFEval) for large language models. IFEval is a straightforward and easy-to-reproduce evaluation benchmark. It focuses on a set of “verifiable instructions” such as “write in more than 400 words” and “mention the keyword of AI at least 3 times”. We identified 25 types of those verifiable instructions and constructed around 500 prompts, with each prompt containing one or more verifiable instructions. We show evaluation results of two widely available LLMs on the market. Our code and data can be found athttps://github.com/google-research/google-research/tree/master/instruction_following_eval
|
2311.10770
|
Language models only really need to use an exponential fraction of their neurons for individual inferences. As proof, we present UltraFastBERT, a BERT variant that uses 0.3% of its neurons during inference while performing on par with similar BERT models.
UltraFastBERT selectively engages just 12 out of 4095 neurons for each layer inference.
This is achieved by replacing feedforward networks with fast feedforward networks (FFFs) . While no truly efficient implementation currently exists to unlock the full acceleration potential of conditional neural execution, we provide high-level CPU code achieving 78x speedup over the optimized baseline feedforward implementation, and a PyTorch implementation delivering 40x speedup over the equivalent batched feedforward inference. We publish our training code, benchmarking setup, and model weights.111https://github.com/pbelcak/UltraFastBERT
|
2308.14711
|
We break the linear link between the layer size and its inference cost by introducing the fast feedforward111https://github.com/pbelcak/fastfeedforward(FFF) architecture, a-time alternative to feedforward networks. We demonstrate that FFFs are up to 220x faster than feedforward networks, up to 6x faster than mixture-of-experts networks, and exhibit better training properties than mixtures of experts thanks to noiseless conditional execution. Pushing FFFs to the limit, we show that they can use as little as 1% of layer neurons for inference in vision transformers while preserving 94.2% of predictive performance.
|
2311.11045
|
Orca 1 learns from rich signals, such as explanation traces, allowing it to outperform conventional instruction-tuned models on benchmarks like BigBench Hard and AGIEval. In Orca 2, we continue exploring how improved training signals can enhance smaller LMs’ reasoning abilities.
Research on training small LMs has often relied on imitation learning to replicate the output of more capable models. We contend that excessive emphasis on imitation may restrict the potential of smaller models. We seek to teach small LMs to employ different solution strategies for different tasks, potentially different from the one used by the larger model. For example, while larger models might provide a direct answer to a complex task, smaller models may not have the same capacity.
In Orca 2, we teach the model various reasoning techniques (step-by-step, recall then generate, recall-reason-generate, direct answer, etc.) . Moreover, we aim to help the model learn to determine the most effective solution strategy for each task. We evaluate Orca 2 using a comprehensive set of 15 diverse benchmarks (corresponding to approximately 100 tasks and over 36K unique prompts) . Orca 2 significantly surpasses models of similar size and attains performance levels similar or better to those of models 5-10x larger, as assessed on complex tasks that test advanced reasoning abilities in zero-shot settings. We make Orca 2 weights publicly available ataka.ms/orca-lmto support research on the development, evaluation, and alignment of smaller LMs.
|
2310.00035
|
Fine-tuned LLMs often exhibit poor uncertainty quantification, manifesting as overconfidence, poor calibration, and unreliable prediction results on test data or out-of-distribution samples.
One approach commonly used in vision for alleviating this issue is a deep ensemble, which constructs an ensemble by training the same model multiple times using different random initializations.
However, there is a huge challenge to ensembling LLMs: the most effective LLMs are very, very large.
Keeping a single LLM in memory is already challenging enough: keeping an ensemble of e.g. 5 LLMs in memory is impossible in many settings.
To address this issue, we propose an ensemble approach using Low-Rank Adapters (LoRA) , a parameter-efficient fine-tuning technique.
Critically, these low-rank adapters require a very small number of parameters, orders of magnitude less than the underlying pre-trained model.
Thus, it is possible to construct large ensembles of LoRA adapters with almost the same computational overhead as using the original model.
We find that LoRA ensembles, applied on its own or on top of pre-existing regularization techniques, gives consistent improvements in predictive accuracy and uncertainty quantification.
|
2309.02784
|
As the size of large language models (LLMs) continues to grow, model compression without sacrificing accuracy has become a crucial challenge for deployment.
While some quantization methods, such as GPTQ, have made progress in achieving acceptable 4-bit weight-only quantization, attempts at lower bit quantization often result in severe performance degradation.
In this paper, we introduce a technique called norm tweaking, which can be used as a plugin in current PTQ methods to achieve high precision while being cost-efficient. Our approach is inspired by the observation that rectifying the quantized activation distribution to match its float counterpart can readily restore accuracy for LLMs.
To achieve this, we carefully design a tweaking strategy that includes calibration data generation and channel-wise distance constraint to update the weights of normalization layers for better generalization. We conduct extensive experiments on various datasets using several open-sourced LLMs. Our method demonstrates significant improvements in both weight-only quantization and joint quantization of weights and activations, surpassing existing PTQ methods.
On GLM-130B and OPT-66B, our method even achieves the same level of accuracy at 2-bit quantization as their float ones. Our simple and effective approach makes it more practical for real-world applications.
|
2109.10086
|
In neural Information Retrieval (IR) , ongoing research is directed towards improving the first retriever in ranking pipelines. Learning dense embeddings to conduct retrieval using efficient approximate nearest neighbors methods has proven to work well. Meanwhile, there has been a growing interest in learningsparserepresentations for documents and queries, that could inherit from the desirable properties of bag-of-words models such as the exact matching of terms and the efficiency of inverted indexes. Introduced recently, the SPLADE model provides highly sparse representations and competitive results with respect to state-of-the-art dense and sparse approaches. In this paper, we build on SPLADE and propose several significant improvements in terms of effectiveness and/or efficiency. More specifically, we modify the pooling mechanism, benchmark a model solely based on document expansion, and introduce models trained with distillation. We also report results on the BEIR benchmark. Overall, SPLADE is considerably improved with more than% gains on NDCG@10 on TREC DL 2019, leading to state-of-the-art results on the BEIR benchmark.
|
2107.05720
|
In neural Information Retrieval, ongoing research is directed towards improving the first retriever in ranking pipelines. Learning dense embeddings to conduct retrieval using efficient approximate nearest neighbors methods has proven to work well. Meanwhile, there has been a growing interest in learningsparserepresentations for documents and queries, that could inherit from the desirable properties of bag-of-words models such as the exact matching of terms and the efficiency of inverted indexes. In this work, we present a new first-stage ranker based on explicit sparsity regularization and a log-saturation effect on term weights, leading to highly sparse representations and competitive results with respect to state-of-the-art dense and sparse methods. Our approach is simple, trained end-to-end in a single stage. We also explore the trade-off between effectiveness and efficiency, by controlling the contribution of the sparsity regularization.
|
2212.10670
|
Given the success with in-context learning of large pre-trained language models, we introduce in-context learning distillation to transfer in-context few-shot learning ability from large models to smaller models.
We propose to combine in-context learning objectives with language modeling objectives to distill both the ability to read in-context examples and task knowledge to the smaller models. We perform in-context learning distillation under two different few-shot learning paradigms: Meta In-context Tuning (Meta-ICT) and Multitask In-context Tuning (Multitask-ICT) . Multitask-ICT performs better on multitask few-shot learning but also requires more computation than Meta-ICT.
Our method shows consistent improvements for both Meta-ICT and Multitask-ICT on two benchmarks: LAMA and CrossFit.
Our extensive experiments and analysis reveal that in-context learning objectives and language modeling objectives are complementary under the Multitask-ICT paradigm. In-context learning objectives achieve the best performance when combined with language modeling objectives.
|
2301.12726
|
The surprising ability of Large Language Models (LLMs) to perform well on complex reasoning with only few-shot chain-of-thought prompts is believed to emerge only in very large-scale models (100+ billion parameters) .
We show that such abilities can, in fact, be distilled down from GPT-3.5 (175B) to T5 variants (11B) .
We proposemodel specialization, to specialize the model’s ability towards a target task.
The hypothesis is that
large models (commonly viewed as larger than 100B) have strong modeling power, but are spread on a large spectrum of tasks.
Small models (commonly viewed as smaller than 10B) have limited model capacity, but if we concentrate their capacity on a specific target task,
the model can achieve a decent improved performance.
We use multi-step math reasoning as our testbed because it is a very typical emergent ability.
We show two important aspects of model abilities:
(1) . there exists a very complex balance/ tradeoff between language models’ multi-dimensional abilities;
(2) . by paying the price of decreased generic ability, we can clearly lift up the scaling curve of models smaller than 10B towards a specialized multi-step math reasoning ability.
We further give comprehensive discussions about important design choices for better generalization, including the tuning data format, the start model checkpoint, and a new model selection method.
We hope our practice and
discoveries can serve as an important attempt towards specialized smaller models in the new research paradigm set by LLMs.
|
2205.05055
|
Large transformer-based models are able to perform in-context few-shot learning, without being explicitly trained for it. This observation raises the question: what aspects of the training regime lead to this emergent behavior? Here, we show that this behavior is driven by the distributions of the training data itself. In-context learning emerges when the training data exhibits particular distributional properties such as burstiness (items appear in clusters rather than being uniformly distributed over time) and having large numbers of rarely occurring classes. In-context learning also emerges more strongly when item meanings or interpretations are dynamic rather than fixed. These properties are exemplified by natural language, but are also inherent to naturalistic data in a wide range of other domains. They also depart significantly from the uniform, i.i.d. training distributions typically used for standard supervised learning. In our initial experiments, we found that in-context learning traded off against more conventional weight-based learning, and models were unable to achieve both simultaneously. However, our later experiments uncovered that the two modes of learning could co-exist in a single model when it was trained on data following a skewed Zipfian distribution – another common property of naturalistic data, including language. In further experiments, we found that naturalistic data distributions were only able to elicit in-context learning in transformers, and not in recurrent models. In sum, our findings indicate how the transformer architecture works together with particular properties of the training data to drive the intriguing emergent in-context learning behaviour of large language models, and how future work might encourage both in-context and in-weights learning in domains beyond language.111Code is available at: https://github.com/deepmind/emergent_in_context_learning
|
2311.08360
|
Transformer neural networks can exhibit a surprising capacity for in-context learning (ICL) despite not being explicitly trained for it. Prior work has provided a deeper understanding of how ICL emerges in transformers, e.g., through the lens of mechanistic interpretability, Bayesian inference, or by examining the distributional properties of training data. However, in each of these cases, ICL is treated largely as apersistentphenomenon; namely, once ICL emerges, it is assumed to persist asymptotically. Here, we show that the emergence of ICL during transformer training is, in fact, oftentransient. We train transformers on synthetic data designed so that both ICL and in-weights learning (IWL) strategies can lead to correct predictions. We find that ICL first emerges, then disappears and gives way to IWL, all while the training loss decreases, indicating an asymptotic preference for IWL. The transient nature of ICL is observed in transformers across a range of model sizes and datasets, raising the question of how much to “overtrain” transformers when seeking compact, cheaper-to-run models. We find that L2 regularization may offer a path to more persistent ICL that removes the need for early stopping based on ICL-style validation tasks. Finally, we present initial evidence that ICL transience may be caused by competition between ICL and IWL circuits.
|
2305.14202
|
While large language models (LLMs) can answer many questions correctly, they can also hallucinate and give wrong answers. Wikidata, with its over 12 billion facts, can be used to ground LLMs to improve their factuality. This paper presents WikiWebQuestions, a high-quality question answering benchmark for Wikidata. Ported over from WebQuestions for Freebase, it consists of real-world data with SPARQL annotation. This paper presents a few-shot sequence-to-sequence semantic parser for Wikidata. We modify SPARQL to use the unique domain and property names instead of their IDs. We train the parser to use either the results from an entity linker or mentions in the query. We fine-tune LLaMA by adding the few-shot training data to that used to fine-tune Alpaca. Our experimental results demonstrate the effectiveness of this methodology, establishing a strong baseline of 76% and 65% answer accuracy in the dev and test sets of WikiWebQuestions, respectively. By pairing our semantic parser with GPT-3, we combine verifiable results with qualified GPT-3 guesses to provide useful answers to 96% of the questions in dev. We also show that our method outperforms the state-of-the-art for the QALD-7 Wikidata dataset by 3.6% in F1 score.111Code, data, and model are available athttps://github.com/stanford-oval/wikidata-emnlp23
|
2311.09210
|
Retrieval-augmented language models (RALMs) represent a substantial advancement in the capabilities of large language models, notably in reducing factual hallucination by leveraging external knowledge sources.
However, the reliability of the retrieved information is not always guaranteed. The retrieval of irrelevant data can lead to misguided responses, and potentially causing the model to overlook its inherent knowledge, even when it possesses adequate information to address the query.
Moreover, standard RALMs often struggle to assess whether they possess adequate knowledge, both intrinsic and retrieved, to provide an accurate answer. In situations where knowledge is lacking, these systems should ideally respond with “unknown” when the answer is unattainable.
In response to these challenges, we introducesChain-of-Noting (CoN) , a novel approach aimed at improving the robustness of RALMs in facing noisy, irrelevant documents and in handling unknown scenarios. The core idea ofCoNis to generate sequential reading notes for retrieved documents, enabling a thorough evaluation of their relevance to the given question and integrating this information to formulate the final answer. We employed ChatGPT to create training data forCoN, which was subsequently trained on an LLaMa-2 7B model. Our experiments across four open-domain QA benchmarks show that RALMs equipped withCoNsignificantly outperform standard RALMs.
Notably,CoNachieves an average improvement of +7.9 in EM score given entirely noisy retrieved documents and +10.5 in rejection rates for real-time questions that fall outside the pre-training knowledge scope.
|
2206.14486
|
Widely observed neural scaling laws, in which error falls off as a power of the training set size, model size, or both, have driven substantial performance improvements in deep learning. However, these improvements through scaling alone require considerable costs in compute and energy. Here we focus on the scaling of error with dataset size and show how in theory we can break beyond power law scaling and potentially even reduce it to exponential scaling instead if we have access to a high-quality data pruning metric that ranks the order in which training examples should be discarded to achieve any pruned dataset size. We then test this improved scaling prediction with pruned dataset size empirically, and indeed observe better than power law scaling in practice on ResNets trained on CIFAR-10, SVHN, and ImageNet. Next, given the importance of finding high-quality pruning metrics, we perform the first large-scale benchmarking study of ten different data pruning metrics on ImageNet. We find most existing high performing metrics scale poorly to ImageNet, while the best are computationally intensive and require labels for every image. We therefore developed a new simple, cheap and scalable self-supervised pruning metric that demonstrates comparable performance to the best supervised metrics. Overall, our work suggests that the discovery of good data-pruning metrics may provide a viable path forward to substantially improved neural scaling laws, thereby reducing the resource costs of modern deep learning.
|
2210.14891
|
We present a smoothly broken power law functional form (referred to by us as abroken neural scaling law(BNSL) ) that accurately models andextrapolatesthe scaling behaviors of deep neural networks (i.e. how the evaluation metric of interest varies as amount of compute used for training (or inference) , number of model parameters, training dataset size, model input size, number of training steps, or upstream performance varies) for various architectures and for each of various tasks within a large, diverse set of upstream anddownstreamtasks, in zero-shot, prompted, and fine-tuned settings. This set includes large-scale vision, language, audio, video, diffusion, generative modeling, multimodal learning, contrastive learning, AI alignment, AI capabilities, robotics, out-of-distribution generalization, continual learning, transfer learning, uncertainty estimation / calibration, out-of-distribution detection, adversarial robustness, distillation, sparsity, retrieval, quantization, pruning, fairness, molecules, computer programming/coding, math word problems, “emergent” “phase transitions”, arithmetic, supervised learning, unsupervised/self-supervised learning, and reinforcement learning (single agent and multi-agent) . When compared to other functional forms for neural scaling behavior, this functional form yieldsextrapolationsof scaling behavior that are considerably more accurate on this set. Additionally, this functional form accurately models and extrapolates scaling behavior that other functional forms are incapable of expressing such as the non-monotonic transitions present in the scaling behavior of phenomena such as double descent and the delayed, sharp inflection points present in the scaling behavior of tasks such as arithmetic. Lastly, we use this functional form to glean insights about the limit of the predictability of scaling behavior. Code is available atgithub.com/ethancaballero/broken_neural_scaling_laws
|
2311.09206
|
Semi-structured tables are ubiquitous. There has been a variety of tasks that aim to automatically interpret, augment, and query tables. Current methods often require pretraining on tables or special model architecture design, are restricted to specific table types, or have simplifying assumptions about tables and tasks. This paper makes the first step towards developing open-source large language models (LLMs) as generalists for a diversity of table-based tasks. Towards that end, we constructTableInstruct, a new dataset with a variety of realistic tables and tasks, for instruction tuning and evaluating LLMs. We further develop the first open-source generalist model for tables,TableLlama, by fine-tuning Llama 2 (7B) with LongLoRA to address the long context challenge. We experiment under both in-domain setting and out-of-domain setting. On 7 out of 8 in-domain tasks,TableLlamaachieves comparable or better performance than the SOTA for each task, despite the latter often has task-specific design. On 6 out-of-domain datasets, it achieves 6-48 absolute point gains compared with the base model, showing that training onTableInstructenhances the model’s generalizability. We will open-source our dataset and trained model to boost future work on developing open generalist models for tables.111Work in progress. Code, model and data will be available at:https://osu-nlp-group.github.io/TableLlama/.
|
2012.15828
|
We generalize deep self-attention distillation inMiniLMby only using self-attention relation distillation for task-agnostic compression of pretrained Transformers.
In particular, we define multi-head self-attention relations as scaled dot-product between the pairs ofquery,key, andvaluevectors within each self-attention module.
Then we employ the above relational knowledge to train the student model.
Besides its simplicity and unified principle, more favorably, there is no restriction in terms of the number of student’s attention heads, while most previous work has to guarantee the same head number between teacher and student.
Moreover, the fine-grained self-attention relations tend to fully exploit the interaction knowledge learned by Transformer.
In addition, we thoroughly examine the layer selection strategy for teacher models, rather than just relying on the last layer as inMiniLM.
We conduct extensive experiments on compressing both monolingual and multilingual pretrained models.
Experimental results demonstrate that our models111Distilled models and code will be publicly available athttps://aka.ms/minilm.distilled from base-size and large-size teachers (BERT, RoBERTa and XLM-R) outperform the state-of-the-art.
|
2002.10957
|
Pre-trained language models (e.g., BERTand its variants) have achieved remarkable success in varieties of NLP tasks.
However, these models usually consist of hundreds of millions of parameters which brings challenges for fine-tuning and online serving in real-life applications
due to latency and capacity constraints.
In this work, we present a simple and effective approach to compress large Transformerbased pre-trained models, termed as deep self-attention distillation.
The small model (student) is trained by deeply mimicking the self-attention module, which plays a vital role in Transformer networks, of the large model (teacher) .
Specifically, we propose distilling the self-attention module of the last Transformer layer of the teacher, which is effective and flexible for the student.
Furthermore, we introduce the scaled dot-product between values in the self-attention module as the new deep self-attention knowledge, in addition to the attention distributions (i.e., the scaled dot-product of queries and keys) that have been used in existing works.
Moreover, we show that introducing a teacher assistantalso helps the distillation of large pre-trained Transformer models.
Experimental results demonstrate that our monolingual model111The code and models are publicly available athttps://aka.ms/minilm.outperforms state-of-the-art baselines in different parameter size of student models.
In particular, it retains more thanaccuracy on SQuAD 2.0 and several GLUE benchmark tasks usingof the Transformer parameters and computations of the teacher model.
We also obtain competitive results in applying deep self-attention distillation to multilingual pre-trained models.
|
2305.11627
|
Large language models (LLMs) have shown remarkable capabilities in language understanding and generation. However, such impressive capability typically comes with a substantial model size, which presents significant challenges in both the deployment, inference, and training stages. With LLM being a general-purpose task solver, we explore its compression in a task-agnostic manner, which aims to preserve the multi-task solving and language generation ability of the original LLM. One challenge to achieving this is the enormous size of the training corpus of LLM, which makes both data transfer and model post-training over-burdensome. Thus, we tackle the compression of LLMs within the bound of two constraints: being task-agnostic and minimizing the reliance on the original training dataset. Our method, named LLM-Pruner, adopts structural pruning that selectively removes non-critical coupled structures based on gradient information, maximally preserving the majority of the LLM’s functionality. To this end, the performance of pruned models can be efficiently recovered through tuning techniques, LoRA, in merely3 hours, requiring only50Kdata. We validate the LLM-Pruner on three LLMs, including LLaMA, Vicuna, and ChatGLM, and demonstrate that the compressed models still exhibit satisfactory capabilities in zero-shot classification and generation. The code is available at:https://github.com/horseee/LLM-Pruner
|
2204.00408v3
|
The growing size of neural language models has led to increased attention in model compression.
The two predominant approaches arepruning, which gradually removes weights from a pre-trained model, anddistillation, which trains a smaller compact model to match a larger one.
Pruning methods can significantly reduce the model size but hardly achieve large speedups as distillation. However, distillation methods require large amounts of unlabeled data and are expensive to train.
In this work, we propose a task-specific structured pruning method CoFi111CoFi is pronounced as.(Coarse- andFine-grained Pruning) , which delivers highly parallelizable subnetworks and matches the distillation methods in both accuracy and latency, without resorting to any unlabeled data. Our key insight is to jointly prune coarse-grained (e.g., layers) and fine-grained (e.g., heads and hidden units) modules, which controls the pruning decision of each parameter with masks of different granularity. We also devise a layerwise distillation strategy to transfer knowledge from unpruned to pruned models during optimization. Our experiments on GLUE and SQuAD datasets show that CoFi yields models with over 10speedups with a small accuracy drop, showing its effectiveness and efficiency compared to previous pruning and distillation approaches.222Our code and models are publicly available athttps://github.com/princeton-nlp/CoFiPruning.
|
1708.01009
|
Recurrent neural networks (RNNs) serve as a fundamental building block for many sequence tasks across natural language processing.
Recent research has focused on recurrent dropout techniques or custom RNN cells in order to improve performance.
Both of these can require substantial modifications to the machine learning model or to the underlying RNN configurations.
We revisit traditional regularization techniques, specificallyregularization on RNN activations and slowness regularization over successive hidden states, to improve the performance of RNNs on the task of language modeling.
Both of these techniques require minimal modification to existing RNN architectures and result in performance improvements comparable or superior to more complicated regularization techniques or custom cell architectures.
These regularization techniques can be used without any modification on optimized LSTM implementations such as the NVIDIA cuDNN LSTM.
|
1912.02178
|
Generalization of deep networks has been of great interest in recent years, resulting in a number of theoretically and empirically motivated complexity measures. However, most papers proposing such measures study only a small set of models, leaving open the question of whether the conclusion drawn from those experiments would remain valid in other settings. We present the first large scale study of generalization in deep networks. We investigate more then 40 complexity measures taken from both theoretical bounds and empirical studies. We train over 10,000 convolutional networks by systematically varying commonly used hyperparameters. Hoping to uncover potentially causal relationships between each measure and generalization, we analyze carefully controlled experiments and show surprising failures of some measures as well as promising measures for further research.
|
2311.06243
|
Large foundation models are becoming ubiquitous, but training them from scratch is prohibitively expensive.
Thus, efficiently adapting these powerful models to downstream tasks is increasingly important. In this paper, we study a principled finetuning paradigm – Orthogonal Finetuning (OFT) – for downstream task adaptation. Despite demonstrating good generalizability, OFT still uses a fairly large number of trainable parameters due to the high dimensionality of orthogonal matrices. To address this, we start by examining OFT from an information transmission perspective, and then identify a few key desiderata that enable better parameter-efficiency. Inspired by how the Cooley-Tukey fast Fourier transform algorithm enables efficient information transmission, we propose an efficient orthogonal parameterization using butterfly structures. We apply this parameterization to OFT, creating a novel parameter-efficient finetuning method, called Orthogonal Butterfly (BOFT) . By subsuming OFT as a special case, BOFT introduces a generalized orthogonal finetuning framework. Finally, we conduct an extensive empirical study of adapting large vision transformers, large language models, and text-to-image diffusion models to various downstream tasks in vision and language.
|
2311.05657
|
We introduceLumos, a novel framework for traininglanguage agents that employs aunified data format and amodular architecture based on open-source large language models (LLMs) .Lumosconsists of three distinct modules: planning, grounding, and execution. Theplanning modulebreaks down a task into a series of high-level, tool-agnostic subgoals, which are then made specific by thegrounding modulethrough a set of low-level actions. These actions are subsequently executed by theexecution module, utilizing a range of off-the-shelf tools and APIs. In order to train these modules effectively, high-quality annotations of subgoals and actions were collected and are made available for fine-tuning open-source LLMs for various tasks such as complex question answering, web tasks, and math problems. Leveraging this unified data and modular design,Lumosnot only achieves comparable or superior performance to current, state-of-the-art agents, but also exhibits several key advantages: (1) Lumossurpasses GPT-4/3.5-based agents in complex question answering and web tasks, while equalling the performance of significantly larger LLM agents on math tasks; (2) Lumosoutperforms open-source agents created through conventional training methods and those using chain-of-thoughts training; and (3) Lumosis capable of effectively generalizing to unseen interactive tasks, outperforming larger LLM-based agents and even exceeding performance of specialized agents.
|
2310.01208
|
The recent success of Large Language Models (LLMs) has gained significant attention in both academia and industry. Substantial efforts have been made to enhance the zero- and few-shot generalization capabilities of open-source LLMs through finetuning. Currently, the prevailing approach is instruction-tuning, which trains LLMs to complete real-world tasks by generating responses guided by natural language instructions. It is worth noticing that such an approach may underperform in sequence and token classification tasks. Unlike text generation tasks, classification tasks have a limited label space, where precise label prediction is more appreciated than generating diverse and human-like responses. Prior research has unveiled that instruction-tuned LLMs cannot outperform BERT, prompting us to explore the potential of leveraging latent representations from LLMs for supervised label prediction. In this paper, we introduce a label-supervised adaptation for LLMs, which aims to finetuning the model with discriminant labels. We evaluate this approach with Label Supervised LLaMA (LS-LLaMA) , based on LLaMA-2-7B, a relatively small-scale LLM, and can be finetuned on a single GeForce RTX4090 GPU. We extract latent representations from the final LLaMA layer and project them into the label space to compute the cross-entropy loss. The model is finetuned by Low-Rank Adaptation (LoRA) to minimize this loss. Remarkably, without intricate prompt engineering or external knowledge, LS-LLaMA substantially outperforms LLMs ten times its size in scale and demonstrates consistent improvements compared to robust baselines like BERT-Large and RoBERTa-Large in text classification. Moreover, by removing the causal mask from decoders, LS-unLLaMA achieves the state-of-the-art performance in named entity recognition (NER) . Our work will shed light on a novel approach to adapting LLMs for various downstream tasks.
|
2305.12129
|
Pretrained language models (LMs) have shown compelling performance on various downstream tasks, but unfortunately they require a tremendous amount of inference compute. Knowledge distillation finds a path to compress LMs to small ones with a teacher-student paradigm. However, when the capacity gap between the teacher and the student is large, a curse of capacity gap appears, invoking a deficiency in distilling LMs. While a few studies have been carried out to fill the gap, the curse is not yet well tackled. In this paper, we aim at lifting the curse of capacity gap via enlarging the capacity of the student without notably increasing the inference compute. Largely motivated by sparse activation regime of mixture of experts (MoE) , we propose a mixture of minimal experts (MiniMoE) , which imposes extra parameters to the student but introduces almost no additional inference compute. Experimental results on GLUE and CoNLL demonstrate the curse of capacity gap is lifted by the magic ofMiniMoEto a large extent.MiniMoEalso achieves the state-of-the-art performance at small FLOPs compared with a range of competitive baselines. With a compression rate as much as50,MiniMoEpreserves95% GLUE score of the teacher.111Code is available athttps://github.com/GeneZC/MiniMoE
|
1806.03377
|
PipeDream is a Deep Neural Network (DNN) training system for GPUs
that parallelizes computation by pipelining execution across multiple machines.
Itspipeline parallelcomputing model avoids the slowdowns faced
by data-parallel training when large models and/or limited
network bandwidth induce high communication-to-computation ratios.
PipeDream reduces communication by up to 95% for
large DNNs relative to data-parallel training, and
allows perfect overlap of communication and computation.
PipeDream keeps all available GPUs productive by systematically
partitioning DNN layers among them to balance work and minimize communication,
versions model parameters for backward pass correctness, and
schedules the forward and backward passes of different inputs
in round-robin fashion to optimize “time to target accuracy”.
Experiments with five different DNNs on two different clusters
show that PipeDream is up to 5x faster in time-to-accuracy
compared to data-parallel training.
|
1812.06162
|
In an increasing number of domains it has been demonstrated that deep learning models can be trained using relatively large batch sizes without sacrificing data efficiency. However the limits of this massive data parallelism seem to differ from domain to domain, ranging from batches of tens of thousands in ImageNet to batches of millions in RL agents that play the game Dota 2.
To our knowledge there is limited conceptual understanding of why these limits to batch size differ or how we might choose the correct batch size in a new domain.
In this paper, we demonstrate that a simple and easy-to-measure statistic called thegradient noise scalepredicts the largest useful batch size across many domains and applications, including a number of supervised learning datasets (MNIST, SVHN, CIFAR-10, ImageNet, Billion Word) , reinforcement learning domains (Atari and Dota) , and even generative model training (autoencoders on SVHN) . We find that the noise scale increases as the loss decreases over a training run and depends on the model size primarily through improved model performance. Our empirically-motivated theory also describes the tradeoff between compute-efficiency and time-efficiency, and provides a rough model of the benefits of adaptive batch-size training.
|
2311.02790
|
Citation count of a paper is a commonly used proxy for evaluating the significance of a paper in the scientific community. Yet citation measures are widely criticized for failing to accurately reflect the true impact of a paper.
Thus, we proposeCausalCite, a new way to measure the significance of a paper by assessing the causal impact of the paper on its follow-up papers.CausalCiteis based on a novel causal inference method,TextMatch, which adapts the traditional matching framework to high-dimensional text embeddings.TextMatchencodes each paper using text embeddings from large language models (LLMs) , extracts similar samples by cosine similarity, and synthesizes a counterfactual sample as the weighted average of similar papers according to their similarity values.
We demonstrate the effectiveness ofCausalCiteon various criteria, such as high correlation with paper impact as reported by scientific experts on a previous dataset of 1K papers, (test-of-time) awards for past papers, and its stability across various subfields of AI. We also provide a set of findings that can serve as suggested ways for future researchers to use our metric for a better understanding of the quality of a paper.111Our code is available athttps://github.com/causalNLP/causal-cite.
|
2311.01460
|
To augment language models with the ability to reason, researchers usually prompt or finetune them to produce chain of thought reasoning steps before producing the final answer. However, although people use natural language to reason effectively, it may be that LMs could reason more effectively with some intermediate computation that is not in natural language.
In this work, we explore an alternative reasoning approach: instead of explicitly producing the chain of thought reasoning steps, we use the language model’s internal hidden states to perform implicit reasoning. The implicit reasoning steps are distilled from a teacher model trained on explicit chain-of-thought reasoning, and instead of doing reasoning “horizontally” by producing intermediate words one-by-one, we distill it such that the reasoning happens “vertically” among the hidden states in different layers. We conduct experiments on a multi-digit multiplication task and a grade school math problem dataset and find that this approach enables solving tasks previously not solvable without explicit chain-of-thought, at a speed comparable to no chain-of-thought.
|
2211.05953
|
We introduce Breadth-First Pipeline Parallelism, a novel training schedule which optimizes the combination of pipeline and data parallelism. Breadth-First Pipeline Parallelism lowers training time, cost and memory usage by combining a high GPU utilization with a small batch size per GPU, and by making use of fully sharded data parallelism. Experimentally, we observed an increase of up to 43% in training throughput for a 52 billion-parameter model using a small batch size per GPU compared to Megatron-LM, which would reduce the training time and cost by the same amount on a large GPU cluster.
|
2302.13971
|
We introduce LLaMA, a collection of foundation language models ranging from 7B to 65B parameters.
We train our models on trillions of tokens, and show that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible datasets.
In particular, LLaMA-13B outperforms GPT-3 (175B) on most benchmarks, and LLaMA-65B is competitive with the best models, Chinchilla-70B and PaLM-540B.
We release all our models to the research community111https://github.com/facebookresearch/llama.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.