A Unified Framework for Neural Computation and Learning Over Time

This paper proposes Hamiltonian Learning, a novel unified framework for learning with neural networks "over time", i.e., from a possibly infinite stream of data, in an online manner, without having access to future information. Existing works focus on the simplified setting in which the stream has a known finite length or is segmented into smaller sequences, leveraging well-established learning strategies from statistical machine learning. In this paper, the problem of learning over time is rethought from scratch, leveraging tools from optimal control theory, which yield a unifying view of the temporal dynamics of neural computations and learning. Hamiltonian Learning is based on differential equations that: (i) can be integrated without the need of external software solvers; (ii) generalize the well-established notion of gradient-based learning in feed-forward and recurrent networks; (iii) open to novel perspectives. The proposed framework is showcased by experimentally proving how it can recover gradient-based learning, comparing it to out-of-the box optimizers, and describing how it is flexible enough to switch from fully-local to partially/non-local computational schemes, possibly distributed over multiple devices, and BackPropagation without storing activations. Hamiltonian Learning is easy to implement and can help researches approach in a principled and innovative manner the problem of learning over time.

Continual Learning of Conjugated Visual Representations through Higher-order Motion Flows

Learning with neural networks from a continuous stream of visual information presents several challenges due to the non-i.i.d. nature of the data. However, it also offers novel opportunities to develop representations that are consistent with the information flow. In this paper we investigate the case of unsupervised continual learning of pixel-wise features subject to multiple motion-induced constraints, therefore named motion-conjugated feature representations. Differently from existing approaches, motion is not a given signal (either ground-truth or estimated by external modules), but is the outcome of a progressive and autonomous learning process, occurring at various levels of the feature hierarchy. Multiple motion flows are estimated with neural networks and characterized by different levels of abstractions, spanning from traditional optical flow to other latent signals originating from higher-level features, hence called higher-order motions. Continuously learning to develop consistent multi-order flows and representations is prone to trivial solutions, which we counteract by introducing a self-supervised contrastive loss, spatially-aware and based on flow-induced similarity. We assess our model on photorealistic synthetic streams and real-world videos, comparing to pre-trained state-of-the art feature extractors (also based on Transformers) and to recent unsupervised learning models, significantly outperforming these alternatives.

Dynamic Decoupling of Placid Terminal Attractor-based Gradient Descent Algorithm

Gradient descent (GD) and stochastic gradient descent (SGD) have been widely used in a large number of application domains. Therefore, understanding the dynamics of GD and improving its convergence speed is still of great importance. This paper carefully analyzes the dynamics of GD based on the terminal attractor at different stages of its gradient flow. On the basis of the terminal sliding mode theory and the terminal attractor theory, four adaptive learning rates are designed. Their performances are investigated in light of a detailed theoretical investigation, and the running times of the learning procedures are evaluated and compared. The total times of their learning processes are also studied in detail. To evaluate their effectiveness, various simulation results are investigated on a function approximation problem and an image classification problem.

State-Space Modeling in Long Sequence Processing: A Survey on Recurrence in the Transformer Era

Effectively learning from sequential data is a longstanding goal of Artificial Intelligence, especially in the case of long sequences. From the dawn of Machine Learning, several researchers engaged in the search of algorithms and architectures capable of processing sequences of patterns, retaining information about the past inputs while still leveraging the upcoming data, without losing precious long-term dependencies and correlations. While such an ultimate goal is inspired by the human hallmark of continuous real-time processing of sensory information, several solutions simplified the learning paradigm by artificially limiting the processed context or dealing with sequences of limited length, given in advance. These solutions were further emphasized by the large ubiquity of Transformers, that have initially shaded the role of Recurrent Neural Nets. However, recurrent networks are facing a strong recent revival due to the growing popularity of (deep) State-Space models and novel instances of large-context Transformers, which are both based on recurrent computations to go beyond several limits of currently ubiquitous technologies. In fact, the fast development of Large Language Models enhanced the interest in efficient solutions to process data over time. This survey provides an in-depth summary of the latest approaches that are based on recurrent models for sequential data processing. A complete taxonomy over the latest trends in architectural and algorithmic solutions is reported and discussed, guiding researchers in this appealing research field. The emerging picture suggests that there is room for thinking of novel routes, constituted by learning algorithms which depart from the standard Backpropagation Through Time, towards a more realistic scenario where patterns are effectively processed online, leveraging local-forward computations, opening to further research on this topic.

The KANDY Benchmark: Incremental Neuro-Symbolic Learning and Reasoning with Kandinsky Patterns

Artificial intelligence is continuously seeking novel challenges and benchmarks to effectively measure performance and to advance the state-of-the-art. In this paper we introduce KANDY, a benchmarking framework that can be used to generate a variety of learning and reasoning tasks inspired by Kandinsky patterns. By creating curricula of binary classification tasks with increasing complexity and with sparse supervisions, KANDY can be used to implement benchmarks for continual and semi-supervised learning, with a specific focus on symbol compositionality. Classification rules are also provided in the ground truth to enable analysis of interpretable solutions. Together with the benchmark generation pipeline, we release two curricula, an easier and a harder one, that we propose as new challenges for the research community. With a thorough experimental evaluation, we show how both state-of-the-art neural models and purely symbolic approaches struggle with solving most of the tasks, thus calling for the application of advanced neuro-symbolic methods trained over time.

Enhancing Modern Supervised Word Sense Disambiguation Models by Semantic Lexical Resources

Supervised models for Word Sense Disambiguation (WSD) currently yield to state-of-the-art results in the most popular benchmarks. Despite the recent introduction of Word Embeddings and Recurrent Neural Networks to design powerful context-related features, the interest in improving WSD models using Semantic Lexical Resources (SLRs) is mostly restricted to knowledge-based approaches. In this paper, we enhance "modern" supervised WSD models exploiting two popular SLRs: WordNet and WordNet Domains. We propose an effective way to introduce semantic features into the classifiers, and we consider using the SLR structure to augment the training data. We study the effect of different types of semantic features, investigating their interaction with local contexts encoded by means of mixtures of Word Embeddings or Recurrent Neural Networks, and we extend the proposed model into a novel multi-layer architecture for WSD. A detailed experimental comparison in the recent Unified Evaluation Framework (Raganato et al., 2017) shows that the proposed approach leads to supervised models that compare favourably with the state-of-the art.

Neural paraphrasing by automatically crawled and aligned sentence pairs

Paraphrasing is the task of re-writing an input text using other words, without altering the meaning of the original content. Conversational systems can exploit automatic paraphrasing to make the conversation more natural, e.g., talking about a certain topic using different paraphrases in different time instants. Recently, the task of automatically generating paraphrases has been approached in the context of Natural Language Generation (NLG). While many existing systems simply consist in rule-based models, the recent success of the Deep Neural Networks in several NLG tasks naturally suggests the possibility of exploiting such networks for generating paraphrases. However, the main obstacle toward neural-network-based paraphrasing is the lack of large datasets with aligned pairs of sentences and paraphrases, that are needed to efficiently train the neural models. In this paper we present a method for the automatic generation of large aligned corpora, that is based on the assumption that news and blog websites talk about the same events using different narrative styles. We propose a similarity search procedure with linguistic constraints that, given a reference sentence, is able to locate the most similar candidate paraphrases out from millions of indexed sentences. The data generation process is evaluated in the case of the Italian language, performing experiments using pointer-based deep neural architectures.

On the Resurgence of Recurrent Models for Long Sequences -- Survey and Research Opportunities in the Transformer Era

A longstanding challenge for the Machine Learning community is the one of developing models that are capable of processing and learning from very long sequences of data. The outstanding results of Transformers-based networks (e.g., Large Language Models) promotes the idea of parallel attention as the key to succeed in such a challenge, obfuscating the role of classic sequential processing of Recurrent Models. However, in the last few years, researchers who were concerned by the quadratic complexity of self-attention have been proposing a novel wave of neural models, which gets the best from the two worlds, i.e., Transformers and Recurrent Nets. Meanwhile, Deep Space-State Models emerged as robust approaches to function approximation over time, thus opening a new perspective in learning from sequential data, followed by many people in the field and exploited to implement a special class of (linear) Recurrent Neural Networks. This survey is aimed at providing an overview of these trends framed under the unifying umbrella of Recurrence. Moreover, it emphasizes novel research opportunities that become prominent when abandoning the idea of processing long sequences whose length is known-in-advance for the more realistic setting of potentially infinite-length sequences, thus intersecting the field of lifelong-online learning from streamed data.

Neural Time-Reversed Generalized Riccati Equation

Optimal control deals with optimization problems in which variables steer a dynamical system, and its outcome contributes to the objective function. Two classical approaches to solving these problems are Dynamic Programming and the Pontryagin Maximum Principle. In both approaches, Hamiltonian equations offer an interpretation of optimality through auxiliary variables known as costates. However, Hamiltonian equations are rarely used due to their reliance on forward-backward algorithms across the entire temporal domain. This paper introduces a novel neural-based approach to optimal control, with the aim of working forward-in-time. Neural networks are employed not only for implementing state dynamics but also for estimating costate variables. The parameters of the latter network are determined at each time step using a newly introduced local policy referred to as the time-reversed generalized Riccati equation. This policy is inspired by a result discussed in the Linear Quadratic (LQ) problem, which we conjecture stabilizes state dynamics. We support this conjecture by discussing experimental results from a range of optimal control case studies.

Collectionless Artificial Intelligence

By and large, the professional handling of huge data collections is regarded as a fundamental ingredient of the progress of machine learning and of its spectacular results in related disciplines, with a growing agreement on risks connected to the centralization of such data collections. This paper sustains the position that the time has come for thinking of new learning protocols where machines conquer cognitive skills in a truly human-like context centered on environmental interactions. This comes with specific restrictions on the learning protocol according to the collectionless principle, which states that, at each time instant, data acquired from the environment is processed with the purpose of contributing to update the current internal representation of the environment, and that the agent is not given the privilege of recording the temporal stream. Basically, there is neither permission to store the temporal information coming from the sensors, thus promoting the development of self-organized memorization skills at a more abstract level, instead of relying on bare storage to simulate learning dynamics that are typical of offline learning algorithms. This purposely extreme position is intended to stimulate the development of machines that learn to dynamically organize the information by following human-based schemes. The proposition of this challenge suggests developing new foundations on computational processes of learning and reasoning that might open the doors to a truly orthogonal competitive track on AI technologies that avoid data accumulation by design, thus offering a framework which is better suited concerning privacy issues, control and customizability. Finally, pushing towards massively distributed computation, the collectionless approach to AI will likely reduce the concentration of power in companies and governments, thus better facing geopolitical issues.