The machine learning community has become increasingly interested in the energy efficiency of neural networks. The Spiking Neural Network (SNN) is a promising approach to energy-efficient computing, since its activation levels are quantized into temporally sparse, one-bit values (i.e., "spike" events), which additionally converts the sum over weight-activity products into a simple addition of weights (one weight for each spike). However, the goal of maintaining state-of-the-art (SotA) accuracy when converting a non-spiking network into an SNN has remained an elusive challenge, primarily due to spikes having only a single bit of precision. Adopting tools from signal processing, we cast neural activation functions as quantizers with temporally-diffused error, and then train networks while smoothly interpolating between the non-spiking and spiking regimes. We apply this technique to the Legendre Memory Unit (LMU) to obtain the first known example of a hybrid SNN outperforming SotA recurrent architectures---including the LSTM, GRU, and NRU---in accuracy, while reducing activities to at most 3.74 bits on average with 1.26 significant bits multiplying each weight. We discuss how these methods can significantly improve the energy efficiency of neural networks.

Heretofore, neural networks with external memory are restricted to single memory with lossy representations of memory interactions. A rich representation of relationships between memory pieces urges a high-order and segregated relational memory. In this paper, we propose to separate the storage of individual experiences (item memory) and their occurring relationships (relational memory). The idea is implemented through a novel Self-attentive Associative Memory (SAM) operator. Found upon outer product, SAM forms a set of associative memories that represent the hypothetical high-order relationships between arbitrary pairs of memory elements, through which a relational memory is constructed from an item memory. The two memories are wired into a single sequential model capable of both memorization and relational reasoning. We achieve competitive results with our proposed two-memory model in a diversity of machine learning tasks, from challenging synthetic problems to practical testbeds such as geometry, graph, reinforcement learning, and question answering.

This work takes the novel approach of posing the statistical Optimal Transport (OT) problem as that of learning the transport plan's kernel mean embedding. The key advantage is that the estimates for the embeddings of the marginals can now be employed directly, leading to a dimension-free sample complexity for the proposed transport plan and transport map estimators. Also, because of the implicit smoothing in the kernel embeddings, the proposed estimators can perform out-of-sample estimation. Interestingly, the proposed formulation employs an MMD based regularization to avoid overfitting, which is complementary to existing $\phi$-divergence (entropy) based regularization techniques. An appropriate representer theorem is presented that leads to a fully kernelized formulation and hence the same formulation can be used to perform continuous/semi-discrete/discrete OT in any non-standard domain (as long as universal kernels in those domains are known). Finally, an ADMM based algorithm is presented for solving the kernelized formulation efficiently. Empirical results show that the proposed estimator outperforms discrete OT based estimator in terms of transport map accuracy.

Unlike ordinary supervised pattern recognition, in a newly proposed framework namely complementary-label learning, each label specifies one class that the pattern does not belong to. In this paper, we propose the natural generalization of learning from an ordinary label and a complementary label, specifically focused on one-versus-all and pairwise classification. We assume that annotation with a bag of complementary labels is equivalent to providing the rest of all the labels as the candidates of the one true class. Our derived classification risk is in a comprehensive form that includes those in the literature, and succeeded to explicitly show the relationship between the single and multiple ordinary/complementary labels. We further show both theoretically and experimentally that the classification error bound monotonically decreases corresponding to the number of complementary labels. This is consistent because the more complementary labels are provided, the less supervision becomes ambiguous.

Several new properties of weighted Hilbert transform are obtained. If mu is zero, two Plancherel-like equations and the isotropic properties are derived. For mu is real number, a coerciveness is derived and two iterative sequences are constructed to find the inversion. The proposed iterative sequences are applicable to the case of pure imaginary constant mu=i*eta with |eta|

We consider the problem of universal approximation of functions by two-layer neural nets with random weights that are "nearly Gaussian" in the sense of Kullback-Leibler divergence. This problem is motivated by recent works on lazy training, where the weight updates generated by stochastic gradient descent do not move appreciably from the i.i.d. Gaussian initialization. We first consider the mean-field limit, where the finite population of neurons in the hidden layer is replaced by a continual ensemble, and show that our problem can be phrased as global minimization of a free-energy functional on the space of probability measures over the weights. This functional trades off the $L^2$ approximation risk against the KL divergence with respect to a centered Gaussian prior. We characterize the unique global minimizer and then construct a controlled nonlinear dynamics in the space of probability measures over weights that solves a McKean--Vlasov optimal control problem. This control problem is closely related to the Schr\"odinger bridge (or entropic optimal transport) problem, and its value is proportional to the minimum of the free energy. Finally, we show that SGD in the lazy training regime (which can be ensured by jointly tuning the variance of the Gaussian prior and the entropic regularization parameter) serves as a greedy approximation to the optimal McKean--Vlasov distributional dynamics and provide quantitative guarantees on the $L^2$ approximation error.

Machine learning (ML) systems across many application areas are increasingly demonstrating performance that is beyond that of humans. In response to the proliferation of such models, the field of Explainable AI (XAI) has sought to develop techniques that enhance the transparency and interpretability of machine learning methods. In this work, we consider a question not previously explored within the XAI and ML communities: Given a computational system whose performance exceeds that of its human user, can explainable AI capabilities be leveraged to improve the performance of the human? We study this question in the context of the game of Chess, for which computational game engines that surpass the performance of the average player are widely available. We introduce the Rationale-Generating Algorithm, an automated technique for generating rationales for utility-based computational methods, which we evaluate with a multi-day user study against two baselines. The results show that our approach produces rationales that lead to statistically significant improvement in human task performance, demonstrating that rationales automatically generated from an AI's internal task model can be used not only to explain what the system is doing, but also to instruct the user and ultimately improve their task performance.

In this paper we consider infinite horizon discounted dynamic programming problems with finite state and control spaces, and partial state observations. We discuss an algorithm that uses multistep lookahead, truncated rollout with a known base policy, and a terminal cost function approximation. This algorithm is also used for policy improvement in an approximate policy iteration scheme, where successive policies are approximated by using a neural network classifier. A novel feature of our approach is that it is well suited for distributed computation through an extended belief space formulation and the use of a partitioned architecture, which is trained with multiple neural networks. We apply our methods in simulation to a class of sequential repair problems where a robot inspects and repairs a pipeline with potentially several rupture sites under partial information about the state of the pipeline.

We present a map-less path planning algorithm based on Deep Reinforcement Learning (DRL) for mobile robots navigating in unknown environment that only relies on 40-dimensional raw laser data and odometry information. The planner is trained using a reward function shaped based on the online knowledge of the map of the training environment, obtained using grid-based Rao-Blackwellized particle filter, in an attempt to enhance the obstacle awareness of the agent. The agent is trained in a complex simulated environment and evaluated in two unseen ones. We show that the policy trained using the introduced reward function not only outperforms standard reward functions in terms of convergence speed, by a reduction of 36.9\% of the iteration steps, and reduction of the collision samples, but it also drastically improves the behaviour of the agent in unseen environments, respectively by 23\% in a simpler workspace and by 45\% in a more clustered one. Furthermore, the policy trained in the simulation environment can be directly and successfully transferred to the real robot. A video of our experiments can be found at: https://youtu.be/UEV7W6e6ZqI

The ability of humans to quickly identify general concepts from a handful of images has proven difficult to emulate with robots. Recently, a computer architecture was developed that allows robots to mimic some aspects of this human ability by modeling concepts as cognitive programs using an instruction set of primitive cognitive functions. This allowed a robot to emulate human imagination by simulating candidate programs in a world model before generalizing to the physical world. However, this model used a naive search algorithm that required 30 minutes to discover a single concept, and became intractable for programs with more than 20 instructions. To circumvent this bottleneck, we present an algorithm that emulates the human cognitive heuristics of object factorization and sub-goaling, allowing human-level inference speed, improving accuracy, and making the output more explainable.