Communication is a powerful tool for coordination in multi-agent RL. But inducing an effective, common language is a difficult challenge, particularly in the decentralized setting. In this work, we introduce an alternative perspective where communicative messages sent between agents are considered as different incomplete views of the environment state. By examining the relationship between messages sent and received, we propose to learn to communicate using contrastive learning to maximize the mutual information between messages of a given trajectory. In communication-essential environments, our method outperforms previous work in both performance and learning speed. Using qualitative metrics and representation probing, we show that our method induces more symmetric communication and captures global state information from the environment. Overall, we show the power of contrastive learning and the importance of leveraging messages as encodings for effective communication.

For communication to happen successfully, a common language is required between agents to understand information communicated by one another. Inducing the emergence of a common language has been a difficult challenge to multi-agent learning systems. In this work, we introduce an alternative perspective to the communicative messages sent between agents, considering them as different incomplete views of the environment state. Based on this perspective, we propose a simple approach to induce the emergence of a common language by maximizing the mutual information between messages of a given trajectory in a self-supervised manner. By evaluating our method in communication-essential environments, we empirically show how our method leads to better learning performance and speed, and learns a more consistent common language than existing methods, without introducing additional learning parameters.

We propose SwitchTT, a multi-task extension to Trajectory Transformer but enhanced with two striking features: (i) exploiting a sparsely activated model to reduce computation cost in multi-task offline model learning and (ii) adopting a distributional trajectory value estimator that improves policy performance, especially in sparse reward settings. These two enhancements make SwitchTT suitable for solving multi-task offline reinforcement learning problems, where model capacity is critical for absorbing the vast quantities of knowledge available in the multi-task dataset. More specifically, SwitchTT exploits switch transformer model architecture for multi-task policy learning, allowing us to improve model capacity without proportional computation cost. Also, SwitchTT approximates the distribution rather than the expectation of trajectory value, mitigating the effects of the Monte-Carlo Value estimator suffering from poor sample complexity, especially in the sparse-reward setting. We evaluate our method using the suite of ten sparse-reward tasks from the gym-mini-grid environment.We show an improvement of 10% over Trajectory Transformer across 10-task learning and obtain up to 90% increase in offline model training speed. Our results also demonstrate the advantage of the switch transformer model for absorbing expert knowledge and the importance of value distribution in evaluating the trajectory.

We study two state-of-the-art solutions to the multi-agent pickup and delivery (MAPD) problem based on different principles -- multi-agent path-finding (MAPF) and multi-agent reinforcement learning (MARL). Specifically, a recent MAPF algorithm called conflict-based search (CBS) and a current MARL algorithm called shared experience actor-critic (SEAC) are studied. While the performance of these algorithms is measured using quite different metrics in their separate lines of work, we aim to benchmark these two methods comprehensively in a simulated warehouse automation environment.

Recent techniques in dynamical scheduling and resource management have found applications in warehouse environments due to their ability to organize and prioritize tasks in a higher temporal resolution. The rise of deep reinforcement learning, as a learning paradigm, has enabled decentralized agent populations to discover complex coordination strategies. However, training multiple agents simultaneously introduce many obstacles in training as observation and action spaces become exponentially large. In our work, we experimentally quantify how various aspects of the warehouse environment (e.g., floor plan complexity, information about agents' live location, level of task parallelizability) affect performance and execution priority. To achieve efficiency, we propose a compact representation of the state and action space for location-aware multi-agent systems, wherein each agent has knowledge of only self and task coordinates, hence only partial observability of the underlying Markov Decision Process. Finally, we show how agents trained in certain environments maintain performance in completely unseen settings and also correlate performance degradation with floor plan geometry.

In a warehouse environment, tasks appear dynamically. Consequently, a task management system that matches them with the workforce too early (e.g., weeks in advance) is necessarily sub-optimal. Also, the rapidly increasing size of the action space of such a system consists of a significant problem for traditional schedulers. Reinforcement learning, however, is suited to deal with issues requiring making sequential decisions towards a long-term, often remote, goal. In this work, we set ourselves on a problem that presents itself with a hierarchical structure: the task-scheduling, by a centralised agent, in a dynamic warehouse multi-agent environment and the execution of one such schedule, by decentralised agents with only partial observability thereof. We propose to use deep reinforcement learning to solve both the high-level scheduling problem and the low-level multi-agent problem of schedule execution. Finally, we also conceive the case where centralisation is impossible at test time and workers must learn how to cooperate in executing the tasks in an environment with no schedule and only partial observability.

In multi-agent collaboration problems with communication, an agent's ability to encode their intention and interpret other agents' strategies is critical for planning their future actions. This paper introduces a novel algorithm called Intention Embedded Communication (IEC) to mimic an agent's language learning ability. IEC contains a perception module for decoding other agents' intentions in response to their past actions. It also includes a language generation module for learning implicit grammar during communication with two or more agents. Such grammar, by construction, should be compact for efficient communication. Both modules undergo conjoint evolution - similar to an infant's babbling that enables it to learn a language of choice by trial and error. We utilised three multi-agent environments, namely predator/prey, traffic junction and level-based foraging and illustrate that such a co-evolution enables us to learn much quicker (50%) than state-of-the-art algorithms like MADDPG. Ablation studies further show that disabling the inferring belief module, communication module, and the hidden states reduces the model performance by 38%, 60% and 30%, respectively. Hence, we suggest that modelling other agents' behaviour accelerates another agent to learn grammar and develop a language to communicate efficiently. We evaluate our method on a set of cooperative scenarios and show its superior performance to other multi-agent baselines. We also demonstrate that it is essential for agents to reason about others' states and learn this ability by continuous communication.

Active inference is a normative framework for explaining behaviour under the free energy principle—a theory of self-organisation originating in neuroscience. It specifies neuronal dynamics for state-estimation in terms of a descent on (variational) free energy—a measure of the fit between an internal (generative) model and sensory observations. The free energy gradient is a prediction error—plausibly encoded in the average membrane potentials of neuronal populations. Conversely, the expected probability of a state can be expressed in terms of neuronal firing rates. We show that this is consistent with current models of neuronal dynamics and establish face validity by synthesising plausible electrophysiological responses. We then show that these neuronal dynamics approximate natural gradient descent, a well-known optimisation algorithm from information geometry that follows the steepest descent of the objective in information space. We compare the information length of belief updating in both schemes, a measure of the distance travelled in information space that has a direct interpretation in terms of metabolic cost. We show that neural dynamics under active inference are metabolically efficient and suggest that neural representations in biological agents may evolve by approximating steepest descent in information space towards the point of optimal inference.

Modern neural networks are highly uncalibrated. It poses a significant challenge for safety-critical systems to utilise deep neural networks (DNNs), reliably. Many recently proposed approaches have demonstrated substantial progress in improving DNN calibration. However, they hardly touch upon refinement, which historically has been an essential aspect of calibration. Refinement indicates separability of a network's correct and incorrect predictions. This paper presents a theoretically and empirically supported exposition for reviewing a model's calibration and refinement. Firstly, we show the breakdown of expected calibration error (ECE), into predicted confidence and refinement. Connecting with this result, we highlight that regularisation based calibration only focuses on naively reducing a model's confidence. This logically has a severe downside to a model's refinement. We support our claims through rigorous empirical evaluations of many state of the art calibration approaches on standard datasets. We find that many calibration approaches with the likes of label smoothing, mixup etc. lower the utility of a DNN by degrading its refinement. Even under natural data shift, this calibration-refinement trade-off holds for the majority of calibration methods. These findings call for an urgent retrospective into some popular pathways taken for modern DNN calibration.

Active inference provides a biologically plausible process theory of brain function. It specifies neuronal dynamics for state-estimation in terms of a gradient descent on (variational) free energy -- a measure of the fit between an internal (generative) model and sensory observations. When formulated for discrete state-space generative models, the free energy gradient turns out to be a prediction error -- plausibly encoded in the average membrane potentials of neuronal populations. Conversely, the expected probability of a state can then be expressed in terms of firing rates. We establish a construct validity to this scheme -- by showing that it is consistent with current models of neuronal dynamics -- and face validity, as it is able to synthesize a wide range of biologically plausible electrophysiological responses. We then show that these neuronal dynamics approximate natural gradient descent, a well-known optimisation algorithm from information geometry that prescribes locally optimal belief updates. Lastly, numerical simulations suggest that both schemes perform equally well on average. The performance of belief updating is scored in terms of information length, a measure of the distance traveled in information space, which has a direct interpretation in terms of metabolic efficiency. These results show that active inference is consistent with state-of-the-art models of neuronal dynamics and coincides with the natural gradient. This suggests that natural selection, by selecting the phenotypes that optimise metabolic and computational efficiency, has implicitly approximated the steepest direction in information space; namely, natural gradient descent.

We propose a novel generative model architecture designed to learn representations for images that factor out a single attribute from the rest of the representation. A single object may have many attributes which when altered do not change the identity of the object itself. Consider the human face; the identity of a particular person is independent of whether or not they happen to be wearing glasses. The attribute of wearing glasses can be changed without changing the identity of the person. However, the ability to manipulate and alter image attributes without altering the object identity is not a trivial task. Here, we are interested in learning a representation of the image that separates the identity of an object (such as a human face) from an attribute (such as 'wearing glasses'). We demonstrate the success of our factorization approach by using the learned representation to synthesize the same face with and without a chosen attribute. We refer to this specific synthesis process as image attribute manipulation. We further demonstrate that our model achieves competitive scores, with state of the art, on a facial attribute classification task.

Image understanding using deep convolutional network has reached human-level performance, yet the closely related problem of video understanding, especially action recognition, has not reached the same required level of maturity. As a solution we propose two independent architectures for action recognition using meta-classifiers – the first is based on combining kernels of support-vector-machines (SVM) and the second is based on distributed Gaussian Processes (GP). Both receive features that are computed using a multi-stream deep convolutional neural network, enabling the achievement of state-of-the-art performance on a 51 and a 101-class action recognition problem (HMDB-51/UCF-101 dataset). We have named the resulting architecture 'pillar networks' as each (very) deep neural network acts as a pillar for the meta-classifiers. In addition, we illustrate that hand-crafted features such as the improved dense trajectories (iDT) and Multi-skip Feature Stacking (MIFS), when used as additional pillars, can further supplement the performance.

The Fisher information metric is an important foundation of information geometry, wherein it allows us to approximate the local geometry of a probability distribution. Recurrent neural networks such as the Sequence-to-Sequence (Seq2Seq) networks that have lately been used to yield state-of-the-art performance on speech translation or image captioning have so far ignored the geometry of the latent embedding, that they iteratively learn. We propose the information geometric Seq2Seq (GeoSeq2Seq) network which abridges the gap between deep recurrent neural networks and information geometry. Specifically, the latent embedding offered by a recurrent network is encoded as a Fisher kernel of a parametric Gaussian Mixture Model, a formalism common in computer vision. We utilise such a network to predict the shortest routes between two nodes of a graph by learning the adjacency matrix using the GeoSeq2Seq formalism; our results show that for such a problem the probabilistic representation of the latent embedding supersedes the non-probabilistic embedding by 10-15\%.

In recent work, it was shown that combining multi-kernel based support vector machines (SVMs) can lead to near state-of-the-art performance on an action recognition dataset (HMDB-51 dataset). In the present work, we show that combining distributed Gaussian Processes with multi-stream deep convolutional neural networks (CNN) alleviate the need to augment a neural network with hand-crafted features. In contrast to prior work, we treat each deep neural convolutional network as an expert wherein the individual predictions (and their respective uncertainties) are combined into a Product of Experts (PoE) framework.

Convolutional and Recurrent, deep neural networks have been successful in machine learning systems for computer vision, reinforcement learning, and other allied fields. However, the robustness of such neural networks is seldom apprised, especially after high classification accuracy has been attained. In this paper, we evaluate the robustness of three recurrent neural networks to tiny perturbations, on three widely used datasets, to argue that high accuracy does not always mean a stable and a robust (to bounded perturbations, adversarial attacks, etc.) system. Especially, normalizing the spectrum of the discrete recurrent network to bound the spectrum (using power method, Rayleigh quotient, etc.) on a unit disk produces stable, albeit highly non-robust neural networks. Furthermore, using the $\epsilon$-pseudo-spectrum, we show that training of recurrent networks, say using gradient-based methods, often result in non-normal matrices that may or may not be diagonalizable. Therefore, the open problem lies in constructing methods that optimize not only for accuracy but also for the stability and the robustness of the underlying neural network, a criterion that is distinct from the other.

Robustness and security of machine learning (ML) systems are intertwined, wherein a non-robust ML system (classifiers, regressors, etc.) can be subject to attacks using a wide variety of exploits. With the advent of scalable deep learning methodologies, a lot of emphasis has been put on the robustness of supervised, unsupervised and reinforcement learning algorithms. Here, we study the robustness of the latent space of a deep variational autoencoder (dVAE), an unsupervised generative framework, to show that it is indeed possible to perturb the latent space, flip the class predictions and keep the classification probability approximately equal before and after an attack. This means that an agent that looks at the outputs of a decoder would remain oblivious to an attack.

Generative adversarial networks (GANs) provide a way to learn deep representations without extensively annotated training data. They achieve this through deriving backpropagation signals through a competitive process involving a pair of networks. The representations that can be learned by GANs may be used in a variety of applications, including image synthesis, semantic image editing, style transfer, image super-resolution and classification. The aim of this review paper is to provide an overview of GANs for the signal processing community, drawing on familiar analogies and concepts where possible. In addition to identifying different methods for training and constructing GANs, we also point to remaining challenges in their theory and application.

Introduction:Attention-deficit hyperactive disorder (ADHD) is the most common neurodevelopmental disorder in children. Diagnosis is currently based on behavioral criteria, but magnetic resonance imaging (MRI) of the brain is increasingly used in ADHD research. To date however, MRI studies have provided mixed results in ADHD patients, particularly with respect to the laterality of findings. Methods:We studied 849 children and adolescents (ages 6-21 y.o.) diagnosed with ADHD (n = 341) and age-matched typically developing (TD) controls with structural brain MRI. We calculated volumetric measures from 34 cortical and 14 non-cortical brain regions per hemisphere, and detailed shape morphometry of subcortical nuclei. Diffusion tensor imaging (DTI) data were collected for a subset of 104 subjects; from these, we calculated mean diffusivity and fractional anisotropy of white matter tracts. Group comparisons were made for within-hemisphere (right/left) and between hemisphere asymmetry indices (AI) for each measure. Results:DTI mean diffusivity AI group differences were significant in cingulum, inferior and superior longitudinal fasciculus, and cortico-spinal tracts (p < 0.001) with the effect of stimulant treatment tending to reduce these patterns of asymmetry differences. Gray matter volumes were more asymmetric in medication free ADHD individuals compared to TD in twelve cortical regions and two non-cortical volumes studied (p < 0.05). Morphometric analyses revealed that caudate, hippocampus, thalamus, and amygdala were more asymmetric (p < 0.0001) in ADHD individuals compared to TD, and that asymmetry differences were more significant than lateralized comparisons. Conclusions:Brain asymmetry measures allow each individual to serve as their own control, diminishing variability between individuals and when pooling data across sites. Asymmetry group differences were more significant than lateralized comparisons between ADHD and TD subjects across morphometric, volumetric, and DTI comparisons.

A widely studied non-polynomial (NP) hard problem lies in finding a route between the two nodes of a graph. Often meta-heuristics algorithms such as $A^{*}$ are employed on graphs with a large number of nodes. Here, we propose a deep recurrent neural network architecture based on the Sequence-2-Sequence model, widely used, for instance in text translation. Particularly, we illustrate that utilising a context vector that has been learned from two different recurrent networks enables increased accuracies in learning the shortest route of a graph. Additionally, we show that one can boost the performance of the Seq2Seq network by smoothing the loss function using a homotopy continuation of the decoder's loss function.

In recent work, it was shown that combining multi-kernel based support vector machines (SVMs) can lead to near state-of-the-art performance on an action recognition dataset (HMDB-51 dataset). This was 0.4\% lower than frameworks that used hand-crafted features in addition to the deep convolutional feature extractors. In the present work, we show that combining distributed Gaussian Processes with multi-stream deep convolutional neural networks (CNN) alleviate the need to augment a neural network with hand-crafted features. In contrast to prior work, we treat each deep neural convolutional network as an expert wherein the individual predictions (and their respective uncertainties) are combined into a Product of Experts (PoE) framework.