Visual cognition of primates is superior to that of artificial neural networks in its ability to 'envision' a visual object, even a newly-introduced one, in different attributes including pose, position, color, texture, etc. To aid neural networks to envision objects with different attributes, we propose a family of objective functions, expressed on groups of examples, as a novel learning framework that we term Group-Supervised Learning (GSL). GSL decomposes inputs into a disentangled representation with swappable components that can be recombined to synthesize new samples, trained through similarity mining within groups of exemplars. For instance, images of red boats & blue cars can be decomposed and recombined to synthesize novel images of red cars. We describe a general class of datasets admissible by GSL. We propose an implementation based on auto-encoder, termed group-supervised zero-shot synthesis network (GZS-Net) trained with our learning framework, that can produce a high-quality red car even if no such example is witnessed during training. We test our model and learning framework on existing benchmarks, in addition to new dataset that we open-source. We qualitatively and quantitatively demonstrate that GZS-Net trained with GSL outperforms state-of-the-art methods

We develop new algorithms for simultaneous learning of multiple tasks (e.g., image classification, depth estimation), and for adapting to unseen task/domain distributions within those high-level tasks (e.g., different environments). First, we learn common representations underlying all tasks. We then propose an attention mechanism to dynamically specialize the network, at runtime, for each task. Our approach is based on weighting each feature map of the backbone network, based on its relevance to a particular task. To achieve this, we enable the attention module to learn task representations during training, which are used to obtain attention weights. Our method improves performance on new, previously unseen environments, and is 1.5x faster than standard existing meta learning methods using similar architectures. We highlight performance improvements for Multi-Task Meta Learning of 4 tasks (image classification, depth, vanishing point, and surface normal estimation), each over 10 to 25 test domains/environments, a result that could not be achieved with standard meta learning techniques like MAML.

Adversarial training, in which a network is trained on both adversarial and clean examples, is one of the most trusted defense methods against adversarial attacks. However, there are three major practical difficulties in implementing and deploying this method - expensive in terms of extra memory and computation costs; accuracy trade-off between clean and adversarial examples; and lack of diversity of adversarial perturbations. Classical adversarial training uses fixed, precomputed perturbations in adversarial examples (input space). In contrast, we introduce dynamic adversarial perturbations into the parameter space of the network, by adding perturbation biases to the fully connected layers of deep convolutional neural network. During training, using only clean images, the perturbation biases are updated in the Fast Gradient Sign Direction to automatically create and store adversarial perturbations by recycling the gradient information computed. The network learns and adjusts itself automatically to these learned adversarial perturbations. Thus, we can achieve adversarial training with negligible cost compared to requiring a training set of adversarial example images. In addition, if combined with classical adversarial training, our perturbation biases can alleviate accuracy trade-off difficulties, and diversify adversarial perturbations.

Intelligent agents can cope with sensory-rich environments by learning task-agnostic state abstractions. In this paper, we propose mechanisms to approximate causal states, which optimally compress the joint history of actions and observations in partially-observable Markov decision processes. Our proposed algorithm extracts causal state representations from RNNs that are trained to predict subsequent observations given the history. We demonstrate that these learned task-agnostic state abstractions can be used to efficiently learn policies for reinforcement learning problems with rich observation spaces. We evaluate agents using multiple partially observable navigation tasks with both discrete (GridWorld) and continuous (VizDoom, ALE) observation processes that cannot be solved by traditional memory-limited methods. Our experiments demonstrate systematic improvement of the DQN and tabular models using approximate causal state representations with respect to recurrent-DQN baselines trained with raw inputs.

Sequential learning of multiple tasks in artificial neural networks using gradient descent leads to catastrophic forgetting, whereby previously learned knowledge is erased during learning of new, disjoint knowledge. Here, we propose a fundamentally new type of method - Beneficial Perturbation Network (BPN). We add task-dependent memory (biasing) units to allow the network to operate in different regimes for different tasks. We compute the most beneficial directions to train these units, in a manner inspired by recent work on adversarial examples. At test time, beneficial perturbations for a given task bias the network toward that task to overcome catastrophic forgetting. BPN is not only more parameter-efficient than network expansion methods, but also does not need to store any data from previous tasks, in contrast with episodic memory methods. Experiments on variants of the MNIST, CIFAR-10, CIFAR-100 datasets demonstrate strong performance of BPN when compared to the state-of-the-art.

Sequential learning of tasks using gradient descent leads to an unremitting decline in the accuracy of tasks for which training data is no longer available, termed catastrophic forgetting. Generative models have been explored as a means to approximate the distribution of old tasks and bypass storage of real data. Here we propose a cumulative closed-loop generator and embedded classifier using an AC-GAN architecture provided with external regularization by a small buffer. We evaluate incremental learning using a notoriously hard paradigm, single headed learning, in which each task is a disjoint subset of classes in the overall dataset, and performance is evaluated on all previous classes. First, we show that the variability contained in a small percentage of a dataset (memory buffer) accounts for a significant portion of the reported accuracy, both in multi-task and continual learning settings. Second, we show that using a generator to continuously output new images while training provides an up-sampling of the buffer, which prevents catastrophic forgetting and yields superior performance when compared to a fixed buffer. We achieve an average accuracy for all classes of 92.26% in MNIST and 76.15% in FASHION-MNIST after 5 tasks using GAN sampling with a buffer of only 0.17% of the entire dataset size. We compare to a network with regularization (EWC) which shows a deteriorated average performance of 29.19% (MNIST) and 26.5% (FASHION). The baseline of no regularization (plain gradient descent) performs at 99.84% (MNIST) and 99.79% (FASHION) for the last task, but below 3% for all previous tasks. Our method has very low long-term memory cost, the buffer, as well as negligible intermediate memory storage.

Sequential learning of multiple tasks in artificial neural networks using gradient descent leads to catastrophic forgetting, whereby previously learned knowledge is erased during learning of new, disjoint knowledge. Here, we propose a new approach to sequential learning which leverages the recent discovery of adversarial examples. We use adversarial subspaces from previous tasks to enable learning of new tasks with less interference. We apply our method to sequentially learning to classify digits 0, 1, 2 (task 1), 4, 5, 6, (task 2), and 7, 8, 9 (task 3) in MNIST (disjoint MNIST task). We compare and combine our Adversarial Direction (AD) method with the recently proposed Elastic Weight Consolidation (EWC) method for sequential learning. We train each task for 20 epochs, which yields good initial performance (99.24% correct task 1 performance). After training task 2, and then task 3, both plain gradient descent (PGD) and EWC largely forget task 1 (task 1 accuracy 32.95% for PGD and 41.02% for EWC), while our combined approach (AD+EWC) still achieves 94.53% correct on task 1. We obtain similar results with a much more difficult disjoint CIFAR10 task, which to our knowledge had not been attempted before (70.10% initial task 1 performance, 67.73% after learning tasks 2 and 3 for AD+EWC, while PGD and EWC both fall to chance level). Our results suggest that AD+EWC can provide better sequential learning performance than either PGD or EWC.

Knowledge distillation (KD) consists of transferring knowledge from one machine learning model (the teacher}) to another (the student). Commonly, the teacher is a high-capacity model with formidable performance, while the student is more compact. By transferring knowledge, one hopes to benefit from the student's compactness. %we desire a compact model with performance close to the teacher's. We study KD from a new perspective: rather than compressing models, we train students parameterized identically to their teachers. Surprisingly, these {Born-Again Networks (BANs), outperform their teachers significantly, both on computer vision and language modeling tasks. Our experiments with BANs based on DenseNets demonstrate state-of-the-art performance on the CIFAR-10 (3.5%) and CIFAR-100 (15.5%) datasets, by validation error. Additional experiments explore two distillation objectives: (i) Confidence-Weighted by Teacher Max (CWTM) and (ii) Dark Knowledge with Permuted Predictions (DKPP). Both methods elucidate the essential components of KD, demonstrating a role of the teacher outputs on both predicted and non-predicted classes. We present experiments with students of various capacities, focusing on the under-explored case where students overpower teachers. Our experiments show significant advantages from transferring knowledge between DenseNets and ResNets in either direction.

Continual Learning in artificial neural networks suffers from interference and forgetting when different tasks are learned sequentially. This paper introduces the Active Long Term Memory Networks (A-LTM), a model of sequential multi-task deep learning that is able to maintain previously learned association between sensory input and behavioral output while acquiring knew knowledge. A-LTM exploits the non-convex nature of deep neural networks and actively maintains knowledge of previously learned, inactive tasks using a distillation loss. Distortions of the learned input-output map are penalized but hidden layers are free to transverse towards new local optima that are more favorable for the multi-task objective. We re-frame the McClelland's seminal Hippocampal theory with respect to Catastrophic Inference (CI) behavior exhibited by modern deep architectures trained with back-propagation and inhomogeneous sampling of latent factors across epochs. We present empirical results of non-trivial CI during continual learning in Deep Linear Networks trained on the same task, in Convolutional Neural Networks when the task shifts from predicting semantic to graphical factors and during domain adaptation from simple to complex environments. We present results of the A-LTM model's ability to maintain viewpoint recognition learned in the highly controlled iLab-20M dataset with 10 object categories and 88 camera viewpoints, while adapting to the unstructured domain of Imagenet with 1,000 object categories.

Many real-world problems have complicated objective functions. To optimize such functions, humans utilize sophisticated sequential decision-making strategies. Many optimization algorithms have also been developed for this same purpose, but how do they compare to humans in terms of both performance and behavior? We try to unravel the general underlying algorithm people may be using while searching for the maximum of an invisible 1D function. Subjects click on a blank screen and are shown the ordinate of the function at each clicked abscissa location. Their task is to find the function’s maximum in as few clicks as possible. Subjects win if they get close enough to the maximum location. Analysis over 23 non-maths undergraduates, optimizing 25 functions from different families, shows that humans outperform 24 well-known optimization algorithms. Bayesian Optimization based on Gaussian Processes, which exploit all the x values tried and all the f(x) values obtained so far to pick the next x, predicts human performance and searched locations better. In 6 follow-up controlled experiments over 76 subjects, covering interpolation, extrapolation, and optimization tasks, we further confirm that Gaussian Processes provide a general and unified theoretical account to explain passive and active function learning and search in humans.