Datasets:

---

taskydata

/

c4tasky_v2

like 0

Follow

Mining Tasky Data 9

ERROR: type should be string, got "https://code.facebook.com/posts/1072626246134461/introducing-fblearner-flow-facebook-s-ai-backbone/ Machine learning automation: Many machine learning algorithms have numerous hyperparameters that can be optimized. At Facebook's scale, a 1 percent improvement in accuracy for many models can have a meaningful impact on people's experiences. So with Flow, we built support for large-scale parameter sweeps and other AutoML features that leverage idle cycles to further improve these models. We are investing further in this area to help scale Facebook's AI/ML experts across many products in Facebook.\nOptunity provides a variety of solvers for hyperparameter tuning problems. A tuning problem is specified by an objective function that provides a score for some tuple of hyperparameters. Specifying the objective function must be done by the user. The software offers a diverse set of solvers to optimize the objective function. A solver determines a good set of hyperparameters.\nA pure learning algorithm can be seen as a function taking training data as input and producing as output a function (e.g. a predictor) or model (i.e. a bunch of functions). However, in practice, many learning algorithms involve hyper-parameters, i.e., annoying knobs to be adjusted. In many algorithms such as Deep Learning algorithms the number of hyper-parameters (ten or more!) can make the idea of having to adjust all of them unappealing. In addition, it has been shown that the use of computer clusters for hyper-parameter selection can have an important effect on results (Pinto et al., 2009). Choosing hyper-parameter values is formally equivalent to the question of model selection, i.e., given a family or set of learning algorithms, how to pick the most appropriate one inside the set? We define a hyper- parameter for a learning algorithm A as a variable to be set prior to the actual application of A to the data, one that is not directly selected by the learning algorithm itself. It is basically an outside control knob. It can be discrete (as in model selection) or continuous (such as the learning rate discussed above). Of course, one can hide these hyper-parameters by wrapping another learning algorithm, say B, around A, to selects A’s hyper-parameters (e.g. to minimize validation set error). We can then call B a hyper-learner, and if B has no hyper-parameters itself then the composition of B over A could be a “pure” learning algorithm, with no hyper-parameter. In the end, to apply a learner to training data, one has to have a pure learning algorithm. The hyper-parameters can be fixed by hand or tuned by an algorithm, but their value has to be selected. The value of some hyperparameters can be selected based on the performance of A on its training data, but most cannot. For any hyper-parameter that has an impact on the effective capacity of a learner, it makes more sense to select its value based on out-of-sample data (outside the training set), e.g., a validation set performance, online error, or cross-validation error. Note that some learning algorithms (in particular unsupervised learning algorithms such as algorithms for training RBMs by approximate maximum likelihood) are problematic in this respect because we cannot directly measure the quantity that is to be optimized (e.g. the likelihood) because it is intractable. On the other hand, the expected denoising reconstruction error is easy to estimate (by just averaging the denoising error over a validation set). Once some out-of-sample data has been used for selecting hyper-parameter values, it cannot be used anymore to obtain an unbiased estimator of generalization performance, so one typically uses a test set (or double cross-validation14, in the case of small datasets) to estimate generalization error of the pure learning algorithm (with hyper-parameter selection hidden inside).\nhttps://code.facebook.com/posts/1072626246134461/introducing-fblearner-flow-facebook-s-ai-backbone/ In some of our earliest work to leverage AI and ML — such as delivering the most relevant content to each person — we noticed that the largest improvements in accuracy often came from quick experiments, feature engineering, and model tuning rather than applying fundamentally different algorithms.\nWe propose an efficient method of ensemble selection for a large ensemble. We prioritize the partitions in the ensemble based on diversity and frequency. Our method selects top K of the partitions in order of priority, where K is decided by the user. We observe that considering jointly the diversity and frequency helps in identifying few representative partitions whose consensus is qualitatively better than the consensus of the entire ensemble.\nTuning hyperparameters is often necessary to get good results with deep neural networks. Typically, the turning is performed either by manual trial-and-error, by using search, or by evaluating validation set performance. The authors propose a gradient based method that is less tedious and less computationally expensive to find good regularization hyperparameters. Unlike previous methods, their method is simpler and computationally lightweight, and it updates both hyperparameters and regular parameters using stochastic gradient descent in the same training run. The gradient of the hyperparameters is obtained from the cost of the unregularized model on the validation set. Although the authors show that their method is effective in finding good regularization hyperparameters, they haven't extended it to common training techniques such as dropout regularization and learning rate adaptation.\nHyperband provides five times to thirty times speedup over state-of-the-art Bayesian optimization algorithms on a variety of deep-learning and kernel-based learning problems.\nThis paper presents a concept of a novel method for adjusting hyper-parameters in Deep Learning (DL) algorithms. An external agent-observer monitors a performance of a selected Deep Learning algorithm. The observer learns to model the DL algorithm using a series of random experiments. Consequently, it may be used for predicting a response of the DL algorithm in terms of a selected quality measurement to a set of hyper-parameters. This allows to construct an ensemble composed of a series of evaluators which constitute an observer-assisted architecture. The architecture may be used to gradually iterate towards to the best achievable quality score in tiny steps governed by a unit of progress. The algorithm is stopped when the maximum number of steps is reached or no further progress is made.\nRecent automated machine learning (AutoML) systems find the right algorithm and hyperparameters in a data-driven way without any human intervention. In this blog post we describe how this is done in our AutoML system Auto-sklearn that won the recent AutoML challenge.\nWe formulate hyperparameter optimization as a pure-exploration non-stochastic infinitely many armed bandit problem where a predefined resource like iterations, data samples, or features is allocated to randomly sampled configurations. We introduce Hyperband for this framework and analyze its theoretical properties, providing several desirable guarantees. Furthermore, we compare Hyperband with state-of-the-art methods on a suite of hyperparameter optimization problems. We observe that Hyperband provides five times to thirty times speedup over state-of-the-art Bayesian optimization algorithms on a variety of deep-learning and kernel-based learning problems.\nThe idea behind the original SuccessiveHalving algorithm follows directly from its name: uniformly allocate a budget to a set of hyperparameter configurations, evaluate the performance of all configurations, throw out the worst half, and repeat until one configurations remains.\nWe study two procedures (reverse-mode and forward-mode) for computing the gradient of the validation error with respect to the hyperparameters of any iterative learning algorithm such as stochastic gradient descent. These procedures mirror two methods of computing gradients for recurrent neural networks and have different trade-offs in terms of running time and space requirements. Our formulation of the reverse-mode procedure is linked to previous work by Maclaurin et al. but does not require reversible dynamics. The forward-mode procedure is suitable for real-time hyperparameter updates, which may significantly speed up hyperparameter optimization on large datasets. We present experiments on data cleaning and on learning task interactions.\nWe propose an extensible and modular language that allows the human expert to compactly represent complex search spaces over architectures and their hyperparameters. The resulting search spaces are tree-structured and therefore easy to traverse. Models can be automatically compiled to computational graphs once values for all hyperparameters have been chosen. We can leverage the structure of the search space to introduce different model search algorithms, such as random search, Monte Carlo tree search (MCTS), and sequential model-based optimization (SMBO).\nAny sufficiently complex system acts as a black box when it becomes easier to experiment with than to understand. Hence, black-box optimization has become increasingly important as systems have become more complex. In this paper we describe Google Vizier, a Google-internal service for performing black-box optimization that has become the de facto parameter tuning engine at Google. Google Vizier is used to optimize many of our machine learning models and other systems, and also provides core capabilities to Google’s Cloud Machine Learning HyperTune subsystem. We discuss our requirements, infrastructure design, underlying algorithms, and advanced features such as transfer learning and automated early stopping that the service provides.\nIn this work, we demonstrate how to learn scalable, convolutional cells from data that transfer to multiple image classification tasks. The key insight to this approach is to design a search space that decouples the complexity of an architecture from the depth of a network. This resulting search space permits identifying good architectures on a small dataset (i.e., CIFAR-10) and transferring the learned architecture to image classifications across a range of data and computational scales. The resulting architectures approach or exceed state-of-the-art performance in terms of CIFAR- 10, ImageNet classification with less computational demand than human-designed architectures. The ImageNet results are particularly important because many state-of-the-art computer vision problems (derive image features or architectures from ImageNet classification models. Finally, we demonstrate that we can employ the resulting learned architecture to perform ImageNet classification with reduced computational budgets that outperform streamlined architectures targeted to mobile and embedded platforms. Our results have strong implications for transfer learning and meta-learning as this is the first work to demonstrate state-of-the-art results using meta-learning on a large scale problem. This work also highlights that learned elements of network architectures, beyond model weights, can transfer across datasets.\n, we use a sequential model-based optimization (SMBO) strategy, in which we search for architectures in order of increasing complexity, while simultaneously learning a surrogate function to guide the search, similar to A* search.\nWe observed that hyperparameter optimization and learning-to-learn share the same mathematical structure, captured by a bilevel programming problem, which consists in minimizing an outer objective function whose value implicitly depends on the solution of an inner problem. The objective function of this latter problem — whose optimization variables are identified with the parameters of (ground) models — is, in turn, parametrized by the outer problem variables, identified either as hyperparameters or parameters of a meta-model, depending on the context. Since the solution of the inner optimization problem does not have, in general, a closed form expression, we formulate a related constrained optimization problem by considering explicitly an optimization dynamics for the inner problem (e.g. gradient descent). In this way we are able to (A) compute the outer objective and optimize it by gradient descent and (B) optimize also variables that parametrize the learning dynamics. We discussed examples of the framework and present experiments on few-shots learning, introducing a method for learning a shared, cross-episode, representation.\nThe algorithm not only searches for the best network topology (e.g., number of layers, type of layers), but also tunes hyper-parameters, such as, learning parameters or data augmentation parameters. The automatic design is achieved using a representation with two distinct levels, where the outer level encodes the general structure of the network, i.e., the sequence of layers, and the inner level encodes the parameters associated with each layer.\nWe formulate the hypernetwork training objective as a compromise between accuracy and diversity, where the diversity takes into account trivial symmetry transformations of the target network.\nWe have developed a new distributed agent IMPALA (Importance-Weighted Actor Learner Architecture) that can scale to thousands of machines and achieve a throughput rate of 250,000 frames per second. We achieve stable learning at high throughput by combining decoupled acting and learning with a novel off-policy correction method called V-trace, which was critical for achieving learning stability. We demonstrate the effectiveness of IMPALA for multi-task reinforcement learning on DMLab-30 (a set of 30 tasks from the DeepMind Lab environment (Beattie et al., 2016)) and Atari-57 (all available Atari games in Arcade Learning Environment (Bellemare et al., 2013a)). Our results show that IMPALA is able to achieve better performance than previous agents, use less data and crucially exhibits positive transfer between tasks as a result of its multi-task approach.\nTo the best of our knowledge, IMPALA is the first Deep-RL agent that has been successfully tested in such large-scale multi-task settings and it has shown superior performance compared to A3C based agents (49.4% vs. 23.8% human normalised score). Most importantly, our experiments on DMLab-30 show that, in the multi-task setting, positive transfer between individual tasks lead IMPALA to achieve better performance compared to the expert training setting. We believe that IMPALA provides a very simple yet scalable and robust framework for building better Deep-RL agents and has the potential to enable research on new challenges.\nUnlike the popular A3C-based agents, in which workers communicate gradients with respect to the parameters of the policy to a central parameter server, IMPALA actors communicate trajectories of experience (sequences of states, actions, and rewards) to a centralised learner.\nWe propose Efficient Neural Architecture Search (ENAS), a fast and inexpensive approach for automatic model design. In ENAS, a controller learns to discover neural network architectures by searching for an optimal subgraph within a large computational graph. The controller is trained with policy gradient to select a subgraph that maximizes the expected reward on the validation set. Meanwhile the model corresponding to the selected subgraph is trained to minimize a canonical cross entropy loss. Thanks to parameter sharing between child models, ENAS is fast: it delivers strong empirical performances using much fewer GPU-hours than all existing automatic model design approaches, and notably, 1000x less expensive than standard Neural Architecture Search. On the Penn Treebank dataset, ENAS discovers a novel architecture that achieves a test perplexity of 55.8, establishing a new state-of-the-art among all methods without post-training processing. On the CIFAR-10 dataset, ENAS designs novel architectures that achieve a test error of 2.89%, which is on par with NASNet (Zoph et al., 2018), whose test error is 2.65%.\nThe main contribution of this work is to improve the efficiency of NAS by forcing all child models to share weights. The idea has apparent complications, as different child models might utilize their weights differently, but was encouraged by previous work on transfer learning and multitask learning, which have established that parameters learned for a particular model on a particular task can be used for other models on other tasks, with little to no modifications.\nOur results highlight the need for new defenses against our hyperparameter stealing attacks for certain machine learning algorithms.\nOur approach combines a novel hierarchical genetic representation scheme that imitates the modularized design pattern commonly adopted by human experts, and an expressive search space that supports complex topologies. Our algorithm efficiently discovers architectures that outperform a large number of manually designed models for image classification, obtaining top-1 error of 3.6% on CIFAR-10 and 20.3% when transferred to ImageNet, which is competitive with the best existing neural architecture search approaches.\nWe provide an overview of existing work in this field of research and categorize them according to three dimensions: search space, search strategy, and performance estimation strategy.\nOur approach captures the structural similarity between members of a neural network ensemble and train it only once. Subsequently, this knowledge is transferred to all members of the ensemble using function-preserving transformations. Then, these ensemble networks converge significantly faster as compared to training from scratch.\nGHNs model the topology of an architecture and therefore can predict network performance more accurately than regular hypernetworks and premature early stopping. To perform NAS, we randomly sample architectures and use the validation accuracy of networks with GHN generated weights as the surrogate search signal. GHNs are fast – they can search nearly 10 times faster than other random search methods on CIFAR-10 and ImageNet."