--Inherent in virtually every iterative machine learning algorithm is the problem of hyper-parameter tuning which includes three major design parameters: (a) the complexity of the model, e.g., the number of neurons in a neural network, (b) the initial conditions, which heavily affect the behavior of the algorithm, and (c) the dissimilarity measure used to quantify its performance. We introduce an online prototype-based learning algorithm that can be viewed as a progressively growing competitive-learning neural network architecture for classification and clustering. The learning rule of the proposed approach is formulated as an online gradient-free stochastic approximation algorithm that solves a sequence of appropriately defined optimization problems, simulating an annealing process. The annealing nature of the algorithm contributes to avoiding poor local minima, offers robustness with respect to the initial conditions, and provides a means to progressively increase the complexity of the learning model, through an intuitive bifurcation phenomenon. The proposed approach is interpretable, requires minimal hyper-parameter tuning, and allows online control over the performance-complexity trade-off. Finally, we show that Bregman divergences appear naturally as a family of dissimilarity measures that play a central role in both the performance and the computational complexity of the learning algorithm. EARNING from data samples has become an important component of artificial intelligence. While virtually all learning problems can be formulated as constrained stochastic optimization problems, the optimization methods can be intractable, typically dealing with mixed constraints and very large, or even infinite-dimensional spaces [1]. For this reason, feature extraction, model selection and design, and analysis of optimization methods, have been the cornerstone of machine learning algorithms from their genesis until today. Deep learning methods, currently dominating the field of machine learning due to their performance in multiple applications, attempt to learn feature representations from data, using biologically-inspired models in artificial neural networks [2], [3]. Manuscript published in the IEEE Transactions on Neural Networks and Learning Systems (TNNLS).

In this paper we study word stress representations learned by self-supervised speech models (S3M), specifically the Wav2vec 2.0 model. We investigate the S3M representations of word stress for five different languages: Three languages with variable or lexical stress (Dutch, English and German) and two languages with fixed or demarcative stress (Hungarian and Polish). We train diagnostic stress classifiers on S3M embeddings and show that they can distinguish between stressed and unstressed syllables in read-aloud short sentences with high accuracy. We also tested language-specificity effects of S3M word stress. The results indicate that the word stress representations are language-specific, with a greater difference between the set of variable versus the set of fixed stressed languages.

To efficiently factorize high-dimensional distributed representations to the constituent atomic vectors, one can exploit the compute-in-superposition capabilities of vector-symbolic architectures (VSA). Such factorizers however suffer from the phenomenon of limit cycles. Applying noise during the iterative decoding is one mechanism to address this issue. In this paper, we explore ways to further relax the noise requirement by applying noise only at the time of VSA's reconstruction codebook initialization. While the need for noise during iterations proves analog in-memory computing systems to be a natural choice as an implementation media, the adequacy of initialization noise allows digital hardware to remain equally indispensable. This broadens the implementation possibilities of factorizers. Our study finds that while the best performance shifts from initialization noise to iterative noise as the number of factors increases from 2 to 4, both extend the operational capacity by at least 50 times compared to the baseline factorizer resonator networks. Our code is available at: https://github.com/IBM/in-memory-factorizer

--We introduce a real-time identification method for discrete-time state-dependent switching systems in both the input-output and state-space domains. In particular, we design a system of adaptive algorithms running in two timescales; a stochastic approximation algorithm implements an online deterministic annealing scheme at a slow timescale and estimates the mode-switching signal, and an recursive identification algorithm runs at a faster timescale and updates the parameters of the local models based on the estimate of the switching signal. We first focus on piece-wise affine systems and discuss identifiability conditions and convergence properties based on the theory of two-timescale stochastic approximation. In contrast to standard identification algorithms for switched systems, the proposed approach gradually estimates the number of modes and is appropriate for real-time system identification using sequential data acquisition. The progressive nature of the algorithm improves computational efficiency and provides real-time control over the performance-complexity trade-off. Finally, we address specific challenges that arise in the application of the proposed methodology in identification of more general switching systems. Simulation results validate the efficacy of the proposed methodology. Hybrid systems, described by interacting continuous and discrete dynamics, are a powerful modeling tool in the analysis of systems where logic and continuous processes are interlaced, as in most complex cyber-physical systems. In addition to being able to describe switching dynamics, hybrid systems can be used as a tool to approximate highly non-linear dynamics by a collection of simpler models, and boost model explainability and robustness, by decomposing the behavior of a complex system into sub-systems where first principles and domain knowledge can be used for precise model tuning [1], [2]. As a result, hybrid systems have attracted significant attention in the control community. However, first principles modelling is often too complicated and sub-optimal, and a hybrid model needs to be identified on the basis of observations. The majority of the work in this area is based on piece-wise affine (PW A) systems, a class of state-dependent switched systems with important applications in identification, verification, and control synthesis of hybrid and nonlinear systems [2]-[5]. The input-output representation of PW A systems is the class of piece-wise affine auto-regressive exogenous (PW ARX) systems with the switching signal depending on a partitioning of the domain of a vector containing the recent history of input-output pairs.

Distributed sparse block codes (SBCs) exhibit compact representations for encoding and manipulating symbolic data structures using fixed-with vectors. One major challenge however is to disentangle, or factorize, such data structures into their constituent elements without having to search through all possible combinations. This factorization becomes more challenging when queried by noisy SBCs wherein symbol representations are relaxed due to perceptual uncertainty and approximations made when modern neural networks are used to generate the query vectors. To address these challenges, we first propose a fast and highly accurate method for factorizing a more flexible and hence generalized form of SBCs, dubbed GSBCs. Our iterative factorizer introduces a threshold-based nonlinear activation, a conditional random sampling, and an $\ell_\infty$-based similarity metric. Its random sampling mechanism in combination with the search in superposition allows to analytically determine the expected number of decoding iterations, which matches the empirical observations up to the GSBC's bundling capacity. Secondly, the proposed factorizer maintains its high accuracy when queried by noisy product vectors generated using deep convolutional neural networks (CNNs). This facilitates its application in replacing the large fully connected layer (FCL) in CNNs, whereby C trainable class vectors, or attribute combinations, can be implicitly represented by our factorizer having F-factor codebooks, each with $\sqrt[\leftroot{-2}\uproot{2}F]{C}$ fixed codevectors. We provide a methodology to flexibly integrate our factorizer in the classification layer of CNNs with a novel loss function. We demonstrate the feasibility of our method on four deep CNN architectures over CIFAR-100, ImageNet-1K, and RAVEN datasets. In all use cases, the number of parameters and operations are significantly reduced compared to the FCL.

Disentanglement of constituent factors of a sensory signal is central to perception and cognition and hence is a critical task for future artificial intelligence systems. In this paper, we present a compute engine capable of efficiently factorizing holographic perceptual representations by exploiting the computation-in-superposition capability of brain-inspired hyperdimensional computing and the intrinsic stochasticity associated with analog in-memory computing based on nanoscale memristive devices. Such an iterative in-memory factorizer is shown to solve at least five orders of magnitude larger problems that cannot be solved otherwise, while also significantly lowering the computational time and space complexity. We present a large-scale experimental demonstration of the factorizer by employing two in-memory compute chips based on phase-change memristive devices. The dominant matrix-vector multiply operations are executed at O(1) thus reducing the computational time complexity to merely the number of iterations. Moreover, we experimentally demonstrate the ability to factorize visual perceptual representations reliably and efficiently.

In this work, we introduce a learning model designed to meet the needs of applications in which computational resources are limited, and robustness and interpretability are prioritized. Learning problems can be formulated as constrained stochastic optimization problems, with the constraints originating mainly from model assumptions that define a trade-off between complexity and performance. This trade-off is closely related to over-fitting, generalization capacity, and robustness to noise and adversarial attacks, and depends on both the structure and complexity of the model, as well as the properties of the optimization methods used. We develop an online prototype-based learning algorithm based on annealing optimization that is formulated as an online gradient-free stochastic approximation algorithm. The learning model can be viewed as an interpretable and progressively growing competitive-learning neural network model to be used for supervised, unsupervised, and reinforcement learning. The annealing nature of the algorithm contributes to minimal hyper-parameter tuning requirements, poor local minima prevention, and robustness with respect to the initial conditions. At the same time, it provides online control over the performance-complexity trade-off by progressively increasing the complexity of the learning model as needed, through an intuitive bifurcation phenomenon. Finally, the use of stochastic approximation enables the study of the convergence of the learning algorithm through mathematical tools from dynamical systems and control, and allows for its integration with reinforcement learning algorithms, constructing an adaptive state-action aggregation scheme.

The existence of a universal learning architecture in human cognition is a widely spread conjecture supported by experimental findings from neuroscience. While no low-level implementation can be specified yet, an abstract outline of human perception and learning is believed to entail three basic properties: (a) hierarchical attention and processing, (b) memory-based knowledge representation, and (c) progressive learning and knowledge compaction. We approach the design of such a learning architecture from a system-theoretic viewpoint, developing a closed-loop system with three main components: (i) a multi-resolution analysis pre-processor, (ii) a group-invariant feature extractor, and (iii) a progressive knowledge-based learning module. Multi-resolution feedback loops are used for learning, i.e., for adapting the system parameters to online observations. To design (i) and (ii), we build upon the established theory of wavelet-based multi-resolution analysis and the properties of group convolution operators. Regarding (iii), we introduce a novel learning algorithm that constructs progressively growing knowledge representations in multiple resolutions. The proposed algorithm is an extension of the Online Deterministic Annealing (ODA) algorithm based on annealing optimization, solved using gradient-free stochastic approximation. ODA has inherent robustness and regularization properties and provides a means to progressively increase the complexity of the learning model i.e. the number of the neurons, as needed, through an intuitive bifurcation phenomenon. The proposed multi-resolution approach is hierarchical, progressive, knowledge-based, and interpretable. We illustrate the properties of the proposed architecture in the context of the state-of-the-art learning algorithms and deep learning methods.

While deep neural networks (NNs) do not provide the confidence of its prediction, Bayesian neural network (BNN) can estimate the uncertainty of the prediction. However, BNNs have not been widely used in practice due to the computational cost of inference. This prohibitive computational cost is a hindrance especially when processing stream data with low-latency. To address this problem, we propose a novel model which approximate BNNs for data streams. Instead of generating separate prediction for each data sample independently, this model estimates the increments of prediction for a new data sample from the previous predictions. The computational cost of this model is almost the same as that of non-Bayesian NNs. Experiments with semantic segmentation on real-world data show that this model performs significantly faster than BNNs, estimating uncertainty comparable to the results of BNNs.

We describe a type of neural network, called a Resonator Circuit, that factors high-dimensional vectors. Given a composite vector formed by the Hadamard product of several other vectors drawn from a discrete set, a Resonator Circuit can efficiently decompose the composite into these factors. This paper focuses on the case of "bipolar" vectors whose elements are $\pm1$ and characterizes the solution quality, stability properties, and speed of Resonator Circuits in comparison to several benchmark optimization methods including Alternating Least Squares, Iterative Soft Thresholding, and Multiplicative Weights. We find that Resonator Circuits substantially outperform these alternative methods by leveraging a combination of powerful nonlinear dynamics and "searching in superposition", by which we mean that estimates of the correct solution are, at any given time, formed from a weighted superposition of all possible solutions. The considered alternative methods also search in superposition, but the dynamics of Resonator Circuits allow them to strike a more natural balance between exploring the solution space and exploiting local information to drive the network toward probable solutions. Resonator Circuits can be conceptualized as a set of interconnected Hopfield Networks, and this leads to some interesting analysis. In particular, while a Hopfield Network descends an energy function and is guaranteed to converge, a Resonator Circuit is not. However, there exists a high-fidelity regime where Resonator Circuits almost always do converge, and they can solve the factorization problem extremely well. As factorization is central to many aspects of perception and cognition, we believe that Resonator Circuits may bring us a step closer to understanding how this computationally difficult problem is efficiently solved by neural circuits in brains.