Search

Advancements in Large Language Models (LLMs) have opened transformative possibilities for human-robot interaction, especially in collaborative environments. However, Real-time human-AI collaboration requires agents to adapt to unseen human behaviors while maintaining effective communication dynamically. Existing benchmarks fall short in evaluating such adaptability for embodied agents, focusing mostly on the task performance of the agent itself. To address this gap, we propose a novel benchmark that assesses agents' reactive adaptability and instantaneous communication capabilities at every step. Based on this benchmark, we propose a Monitor-then-Adapt framework (MonTA), combining strong adaptability and communication with real-time execution. MonTA contains three key LLM modules, a lightweight \textit{Monitor} for monitoring the need for adaptation in high frequency, and two proficient \textit{Adapters} for subtask and path adaptation reasoning in low frequency. Our results demonstrate that MonTA outperforms other baseline agents on our proposed benchmark. Further user studies confirm the high reasonability adaptation plan and consistent language instruction provided by our framework., Comment: 16 pages, 8 figures

Deep learning architectures based on convolutional neural networks tend to rely on continuous, smooth features. While this characteristics provides significant robustness and proves useful in many real-world tasks, it is strikingly incompatible with the physical characteristic of the world, which, at the scale in which humans operate, comprises crisp objects, typically representing well-defined categories. This study proposes a class of neurosymbolic systems that learn by reconstructing the observed images in terms of visual primitives and are thus forced to form high-level, structural explanations of them. When applied to the task of diagnosing abnormalities in histological imaging, the method proved superior to a conventional deep learning architecture in terms of classification accuracy, while being more transparent., Comment: 16 pages, 9 figures

This work explores expanding the capabilities of large language models (LLMs) pretrained on text to generate 3D meshes within a unified model. This offers key advantages of (1) leveraging spatial knowledge already embedded in LLMs, derived from textual sources like 3D tutorials, and (2) enabling conversational 3D generation and mesh understanding. A primary challenge is effectively tokenizing 3D mesh data into discrete tokens that LLMs can process seamlessly. To address this, we introduce LLaMA-Mesh, a novel approach that represents the vertex coordinates and face definitions of 3D meshes as plain text, allowing direct integration with LLMs without expanding the vocabulary. We construct a supervised fine-tuning (SFT) dataset enabling pretrained LLMs to (1) generate 3D meshes from text prompts, (2) produce interleaved text and 3D mesh outputs as required, and (3) understand and interpret 3D meshes. Our work is the first to demonstrate that LLMs can be fine-tuned to acquire complex spatial knowledge for 3D mesh generation in a text-based format, effectively unifying the 3D and text modalities. LLaMA-Mesh achieves mesh generation quality on par with models trained from scratch while maintaining strong text generation performance., Comment: See the project website at https://research.nvidia.com/labs/toronto-ai/LLaMA-Mesh/

We introduce a multi-turn benchmark for evaluating personalised alignment in LLM-based AI assistants, focusing on their ability to handle user-provided safety-critical contexts. Our assessment of ten leading models across five scenarios (each with 337 use cases) reveals systematic inconsistencies in maintaining user-specific consideration, with even top-rated "harmless" models making recommendations that should be recognised as obviously harmful to the user given the context provided. Key failure modes include inappropriate weighing of conflicting preferences, sycophancy (prioritising user preferences above safety), a lack of attentiveness to critical user information within the context window, and inconsistent application of user-specific knowledge. The same systematic biases were observed in OpenAI's o1, suggesting that strong reasoning capacities do not necessarily transfer to this kind of personalised thinking. We find that prompting LLMs to consider safety-critical context significantly improves performance, unlike a generic 'harmless and helpful' instruction. Based on these findings, we propose research directions for embedding self-reflection capabilities, online user modelling, and dynamic risk assessment in AI assistants. Our work emphasises the need for nuanced, context-aware approaches to alignment in systems designed for persistent human interaction, aiding the development of safe and considerate AI assistants., Comment: Submitted to ICLR 2025 on 01/10/2024

Federated learning (FL) enables collaborative model training across distributed clients while preserving data privacy. Despite its widespread adoption, most FL approaches focusing solely on privacy protection fall short in scenarios where trustworthiness is crucial, necessitating advancements in secure training, dependable decision-making mechanisms, robustness on corruptions, and enhanced performance with Non-IID data. To bridge this gap, we introduce Trustworthy Personalized Federated Learning (TPFL) framework designed for classification tasks via subjective logic in this paper. Specifically, TPFL adopts a unique approach by employing subjective logic to construct federated models, providing probabilistic decisions coupled with an assessment of uncertainty rather than mere probability assignments. By incorporating a trainable heterogeneity prior to the local training phase, TPFL effectively mitigates the adverse effects of data heterogeneity. Model uncertainty and instance uncertainty are further utilized to ensure the safety and reliability of the training and inference stages. Through extensive experiments on widely recognized federated learning benchmarks, we demonstrate that TPFL not only achieves competitive performance compared with advanced methods but also exhibits resilience against prevalent malicious attacks, robustness on domain shifts, and reliability in high-stake scenarios., Comment: 17 Pages with Appendix

Semi-supervised learning (SSL) techniques address the high labeling costs in 3D medical image segmentation, with the teacher-student model being a common approach. However, using an exponential moving average (EMA) in single-teacher models may cause coupling issues, where the weights of the student and teacher models become similar, limiting the teacher's ability to provide additional knowledge for the student. Dual-teacher models were introduced to address this problem but often neglected the importance of maintaining teacher model diversity, leading to coupling issues among teachers. To address the coupling issue, we incorporate a double-copy-paste (DCP) technique to enhance the diversity among the teachers. Additionally, we introduce the Staged Selective Ensemble (SSE) module, which selects different ensemble methods based on the characteristics of the samples and enables more accurate segmentation of label boundaries, thereby improving the quality of pseudo-labels. Experimental results demonstrate the effectiveness of our proposed method in 3D medical image segmentation tasks. Here is the code link: https://github.com/Fazhan-cs/DCP., Comment: 35 pages, 5 figures

Generative Flow Networks (GFlowNets) are amortized inference models designed to sample from unnormalized distributions over composable objects, with applications in generative modeling for tasks in fields such as causal discovery, NLP, and drug discovery. Traditionally, the training procedure for GFlowNets seeks to minimize the expected log-squared difference between a proposal (forward policy) and a target (backward policy) distribution, which enforces certain flow-matching conditions. While this training procedure is closely related to variational inference (VI), directly attempting standard Kullback-Leibler (KL) divergence minimization can lead to proven biased and potentially high-variance estimators. Therefore, we first review four divergence measures, namely, Renyi-$\alpha$'s, Tsallis-$\alpha$'s, reverse and forward KL's, and design statistically efficient estimators for their stochastic gradients in the context of training GFlowNets. Then, we verify that properly minimizing these divergences yields a provably correct and empirically effective training scheme, often leading to significantly faster convergence than previously proposed optimization. To achieve this, we design control variates based on the REINFORCE leave-one-out and score-matching estimators to reduce the variance of the learning objectives' gradients. Our work contributes by narrowing the gap between GFlowNets training and generalized variational approximations, paving the way for algorithmic ideas informed by the divergence minimization viewpoint., Comment: Accepted at NeurIPS 2024, https://openreview.net/forum?id=N5H4z0Pzvn

This paper presents a novel approach to identifying insights of peace in global media using a Retrieval Augmented Generation (RAG) model and concepts of Positive and Negative Intergroup Reciprocity (PIR/NIR). By refining the definitions of PIR and NIR, we offer a more accurate and meaningful analysis of intergroup relations as represented in media articles. Our methodology provides insights into the dynamics that contribute to or detract from peace at a national level., Comment: 6 pages, 1 figure

The goal of this work is to serve as a foundation for deep studies of the topology of state, action, and policy spaces in reinforcement learning. By studying these spaces from a mathematical perspective, we expect to gain more insight into how to build better algorithms to solve decision problems. Therefore, we focus on presenting the connection between the Banach fixed point theorem and the convergence of reinforcement learning algorithms, and we illustrate how the insights gained from this can practically help in designing more efficient algorithms. Before doing so, however, we first introduce relevant concepts such as metric spaces, normed spaces and Banach spaces for better understanding, before expressing the entire reinforcement learning problem in terms of Markov decision processes. This allows us to properly introduce the Banach contraction principle in a language suitable for reinforcement learning, and to write the Bellman equations in terms of operators on Banach spaces to show why reinforcement learning algorithms converge. Finally, we show how the insights gained from the mathematical study of convergence are helpful in reasoning about the best ways to make reinforcement learning algorithms more efficient., Comment: Supervisor : Yae Ulrich Gaba , Mentor : Domini Jocema Leko

We develop a general framework for clustering and distribution matching problems with bandit feedback. We consider a $K$-armed bandit model where some subset of $K$ arms is partitioned into $M$ groups. Within each group, the random variable associated to each arm follows the same distribution on a finite alphabet. At each time step, the decision maker pulls an arm and observes its outcome from the random variable associated to that arm. Subsequent arm pulls depend on the history of arm pulls and their outcomes. The decision maker has no knowledge of the distributions of the arms or the underlying partitions. The task is to devise an online algorithm to learn the underlying partition of arms with the least number of arm pulls on average and with an error probability not exceeding a pre-determined value $\delta$. Several existing problems fall under our general framework, including finding $M$ pairs of arms, odd arm identification, and $M$-ary clustering of $K$ arms belong to our general framework. We derive a non-asymptotic lower bound on the average number of arm pulls for any online algorithm with an error probability not exceeding $\delta$. Furthermore, we develop a computationally-efficient online algorithm based on the Track-and-Stop method and Frank--Wolfe algorithm, and show that the average number of arm pulls of our algorithm asymptotically matches that of the lower bound. Our refined analysis also uncovers a novel bound on the speed at which the average number of arm pulls of our algorithm converges to the fundamental limit as $\delta$ vanishes., Comment: 22 pages, submitted to Information Theory Transactions in September 2024

Consider a collection of data generators which could represent, e.g., humans equipped with a smart-phone or wearables. We want to train a personalized (or tailored) model for each data generator even if they provide only small local datasets. The available local datasets might fail to provide sufficient statistical power to train high-dimensional models (such as deep neural networks) effectively. One possible solution is to identify similar data generators and pool their local datasets to obtain a sufficiently large training set. This paper proposes a novel method for sequentially identifying similar (or relevant) data generators. Our method is similar in spirit to active sampling methods but does not require exchange of raw data. Indeed, our method evaluates the relevance of a data generator by evaluating the effect of a gradient step using its local dataset. This evaluation can be performed in a privacy-friendly fashion without sharing raw data. We extend this method to non-parametric models by a suitable generalization of the gradient step to update a hypothesis using the local dataset provided by a data generator.

Laser absorption spectroscopy (LAS) quantification is a popular tool used in measuring temperature and concentration of gases. It has low error tolerance, whereas current ML-based solutions cannot guarantee their measure reliability. In this work, we propose a new framework, SPEC, to address this issue. In addition to the conventional ML estimator-based estimation mode, SPEC also includes a Physics-driven Anomaly Detection module (PAD) to assess the error of the estimation. And a Correction mode is designed to correct the unreliable estimation. The correction mode is a network-based optimization algorithm, which uses the guidance of error to iteratively correct the estimation. A hybrid surrogate error model is proposed to estimate the error distribution, which contains an ensemble of networks to simulate reconstruction error, and true feasible error computation. A greedy ensemble search is proposed to find the optimal correction robustly and efficiently from the gradient guidance of surrogate model. The proposed SPEC is validated on the test scenarios which are outside the training distribution. The results show that SPEC can significantly improve the estimation quality, and the correction mode outperforms current network-based optimization algorithms. In addition, SPEC has the reconfigurability, which can be easily adapted to different quantification tasks via changing PAD without retraining the ML estimator., Comment: 13 pages

Reward machines allow the definition of rewards for temporally extended tasks and behaviors. Specifying "informative" reward machines can be challenging. One way to address this is to generate reward machines from a high-level abstract description of the learning environment, using techniques such as AI planning. However, previous planning-based approaches generate a reward machine based on a single (sequential or partial-order) plan, and do not allow maximum flexibility to the learning agent. In this paper we propose a new approach to synthesising reward machines which is based on the set of partial order plans for a goal. We prove that learning using such "maximally permissive" reward machines results in higher rewards than learning using RMs based on a single plan. We present experimental results which support our theoretical claims by showing that our approach obtains higher rewards than the single-plan approach in practice., Comment: Paper accepted for publication at the European Conference on Artificial Intelligence (ECAI) 2024

In this paper, we propose an off-policy deep reinforcement learning (DRL) method utilizing the average reward criterion. While most existing DRL methods employ the discounted reward criterion, this can potentially lead to a discrepancy between the training objective and performance metrics in continuing tasks, making the average reward criterion a recommended alternative. We introduce RVI-SAC, an extension of the state-of-the-art off-policy DRL method, Soft Actor-Critic (SAC), to the average reward criterion. Our proposal consists of (1) Critic updates based on RVI Q-learning, (2) Actor updates introduced by the average reward soft policy improvement theorem, and (3) automatic adjustment of Reset Cost enabling the average reward reinforcement learning to be applied to tasks with termination. We apply our method to the Gymnasium's Mujoco tasks, a subset of locomotion tasks, and demonstrate that RVI-SAC shows competitive performance compared to existing methods., Comment: Accepted at ICML 2024; Code: https://github.com/yhisaki/average-reward-drl

The ever-growing speed at which data are generated nowadays, together with the substantial cost of labeling processes cause Machine Learning models to face scenarios in which data are partially labeled. The extreme case where such a supervision is indefinitely unavailable is referred to as extreme verification latency. On the other hand, in streaming setups data flows are affected by exogenous factors that yield non-stationarities in the patterns (concept drift), compelling models learned incrementally from the data streams to adapt their modeled knowledge to the concepts within the stream. In this work we address the casuistry in which these two conditions occur together, by which adaptation mechanisms to accommodate drifts within the stream are challenged by the lack of supervision, requiring further mechanisms to track the evolution of concepts in the absence of verification. To this end we propose a novel approach, AiGAS-dEVL (Adaptive Incremental neural GAS model for drifting Streams under Extreme Verification Latency), which relies on growing neural gas to characterize the distributions of all concepts detected within the stream over time. Our approach exposes that the online analysis of the behavior of these prototypical points over time facilitates the definition of the evolution of concepts in the feature space, the detection of changes in their behavior, and the design of adaptation policies to mitigate the effect of such changes in the model. We assess the performance of AiGAS-dEVL over several synthetic datasets, comparing it to that of state-of-the-art approaches proposed in the recent past to tackle this stream learning setup. Our results reveal that AiGAS-dEVL performs competitively with respect to the rest of baselines, exhibiting a superior adaptability over several datasets in the benchmark while ensuring a simple and interpretable instance-based adaptation strategy., Comment: 26 pages, 3 tables, 4 figures. Under review

Gradient-based optimization has been critical to the success of machine learning, updating a single set of parameters to minimize a single loss. A growing number of applications rely on a generalization of this, where we have a bilevel or nested optimization of which subsets of parameters update on different objectives nested inside each other. We focus on motivating examples of hyperparameter optimization and generative adversarial networks. However, naively applying classical methods often fails when we look at solving these nested problems on a large scale. In this thesis, we build tools for nested optimization that scale to deep learning setups., Comment: View more research details at https://www.jonlorraine.com/

Existing Graph Neural Networks (GNNs) are limited to process graphs each of whose vertices is represented by a vector or a single value, limited their representing capability to describe complex objects. In this paper, we propose the first GNN (called Graph in Graph Neural (GIG) Network) which can process graph-style data (called GIG sample) whose vertices are further represented by graphs. Given a set of graphs or a data sample whose components can be represented by a set of graphs (called multi-graph data sample), our GIG network starts with a GIG sample generation (GSG) module which encodes the input as a \textbf{GIG sample}, where each GIG vertex includes a graph. Then, a set of GIG hidden layers are stacked, with each consisting of: (1) a GIG vertex-level updating (GVU) module that individually updates the graph in every GIG vertex based on its internal information; and (2) a global-level GIG sample updating (GGU) module that updates graphs in all GIG vertices based on their relationships, making the updated GIG vertices become global context-aware. This way, both internal cues within the graph contained in each GIG vertex and the relationships among GIG vertices could be utilized for down-stream tasks. Experimental results demonstrate that our GIG network generalizes well for not only various generic graph analysis tasks but also real-world multi-graph data analysis (e.g., human skeleton video-based action recognition), which achieved the new state-of-the-art results on 13 out of 14 evaluated datasets. Our code is publicly available at https://github.com/wangjs96/Graph-in-Graph-Neural-Network.

When training deep learning models, the performance depends largely on the selected hyperparameters. However, hyperparameter optimization (HPO) is often one of the most expensive parts of model design. Classical HPO methods treat this as a black-box optimization problem. However, gray-box HPO methods, which incorporate more information about the setup, have emerged as a promising direction for more efficient optimization. For example, using intermediate loss evaluations to terminate bad selections. In this work, we propose an HPO method for neural networks using logged checkpoints of the trained weights to guide future hyperparameter selections. Our method, Forecasting Model Search (FMS), embeds weights into a Gaussian process deep kernel surrogate model, using a permutation-invariant graph metanetwork to be data-efficient with the logged network weights. To facilitate reproducibility and further research, we open-source our code at https://github.com/NVlabs/forecasting-model-search., Comment: See the project website at https://research.nvidia.com/labs/toronto-ai/FMS/

The trade-off between expressiveness and interpretability remains a core challenge when building human-centric predictive models for classification and decision-making. While symbolic rules offer interpretability, they often lack expressiveness, whereas neural networks excel in performance but are known for being black boxes. In this paper, we show a combination of Large Language Models (LLMs) and symbolic programs can bridge this gap. In the proposed LLM-based Symbolic Programs (LSPs), the pretrained LLM with natural language prompts provides a massive set of interpretable modules that can transform raw input into natural language concepts. Symbolic programs then integrate these modules into an interpretable decision rule. To train LSPs, we develop a divide-and-conquer approach to incrementally build the program from scratch, where the learning process of each step is guided by LLMs. To evaluate the effectiveness of LSPs in extracting interpretable and accurate knowledge from data, we introduce IL-Bench, a collection of diverse tasks, including both synthetic and real-world scenarios across different modalities. Empirical results demonstrate LSP's superior performance compared to traditional neurosymbolic programs and vanilla automatic prompt tuning methods. Moreover, as the knowledge learned by LSP is a combination of natural language descriptions and symbolic rules, it is easily transferable to humans (interpretable), and other LLMs, and generalizes well to out-of-distribution samples., Comment: Preliminary Version, Code at [this url](https://github.com/ruocwang/llm-symbolic-program)

Recent strides in multimodal model development have ignited a paradigm shift in the realm of text-to-image generation. Among these advancements, CLIP stands out as a remarkable achievement which is a sophisticated autoencoder adept at encoding both textual and visual information within a unified latent space. This paper delves into a comparative analysis between CLIP and its recent counterpart, CLOOB. To unravel the intricate distinctions within the embedding spaces crafted by these models, we employ topological data analysis. Our approach encompasses a comprehensive examination of the modality gap drivers, the clustering structures existing across both high and low dimensions, and the pivotal role that dimension collapse plays in shaping their respective embedding spaces. Empirical experiments substantiate the implications of our analyses on downstream performance across various contextual scenarios. Through this investigation, we aim to shed light on the nuanced intricacies that underlie the comparative efficacy of CLIP and CLOOB, offering insights into their respective strengths and weaknesses, and providing a foundation for further refinement and advancement in multimodal model research., Comment: 10 pages, 17 figures, 2 tables

Traditionally, cognitive and computer scientists have viewed intelligence solipsistically, as a property of unitary agents devoid of social context. Given the success of contemporary learning algorithms, we argue that the bottleneck in artificial intelligence (AI) progress is shifting from data assimilation to novel data generation. We bring together evidence showing that natural intelligence emerges at multiple scales in networks of interacting agents via collective living, social relationships and major evolutionary transitions, which contribute to novel data generation through mechanisms such as population pressures, arms races, Machiavellian selection, social learning and cumulative culture. Many breakthroughs in AI exploit some of these processes, from multi-agent structures enabling algorithms to master complex games like Capture-The-Flag and StarCraft II, to strategic communication in Diplomacy and the shaping of AI data streams by other AIs. Moving beyond a solipsistic view of agency to integrate these mechanisms suggests a path to human-like compounding innovation through ongoing novel data generation., Comment: 17 pages, 2 figures, 1 box

Machine learning (ML) models are increasingly used in various applications, from recommendation systems in e-commerce to diagnosis prediction in healthcare. In this paper, we present a novel dynamic framework for thinking about the deployment of ML models in a performative, human-ML collaborative system. In our framework, the introduction of ML recommendations changes the data-generating process of human decisions, which are only a proxy to the ground truth and which are then used to train future versions of the model. We show that this dynamic process in principle can converge to different stable points, i.e. where the ML model and the Human+ML system have the same performance. Some of these stable points are suboptimal with respect to the actual ground truth. As a proof of concept, we conduct an empirical user study with 1,408 participants. In the study, humans solve instances of the knapsack problem with the help of machine learning predictions of varying performance. This is an ideal setting because we can identify the actual ground truth, and evaluate the performance of human decisions supported by ML recommendations. We find that for many levels of ML performance, humans can improve upon the ML predictions. We also find that the improvement could be even higher if humans rationally followed the ML recommendations. Finally, we test whether monetary incentives can increase the quality of human decisions, but we fail to find any positive effect. Using our empirical data to approximate our collaborative system suggests that the learning process would dynamically reach an equilibrium performance that is around 92% of the maximum knapsack value. Our results have practical implications for the deployment of ML models in contexts where human decisions may deviate from the indisputable ground truth., Comment: 10 Pages and appendix

In this paper, we show how mixed-integer conic optimization can be used to combine feature subset selection with holistic generalized linear models to fully automate the model selection process. Concretely, we directly optimize for the Akaike and Bayesian information criteria while imposing constraints designed to deal with multicollinearity in the feature selection task. Specifically, we propose a novel pairwise correlation constraint that combines the sign coherence constraint with ideas from classical statistical models like Ridge regression and the OSCAR model.

Pareto front learning is a technique that introduces preference vectors in a neural network to approximate the Pareto front. Previous Pareto front learning methods have demonstrated high performance in approximating simple Pareto fronts. These methods often sample preference vectors from a fixed Dirichlet distribution. However, no fixed sampling distribution can be adapted to diverse Pareto fronts. Efficiently sampling preference vectors and accurately estimating the Pareto front is a challenge. To address this challenge, we propose a data-driven preference vector sampling framework for Pareto front learning. We utilize the posterior information of the objective functions to adjust the parameters of the sampling distribution flexibly. In this manner, the proposed method can sample preference vectors from the location of the Pareto front with a high probability. Moreover, we design the distribution of the preference vector as a mixture of Dirichlet distributions to improve the performance of the model in disconnected Pareto fronts. Extensive experiments validate the superiority of the proposed method compared with state-of-the-art algorithms., Comment: International Joint Conference on Neural Network (IJCNN'24)

Automatic workflow composition (AWC) is a relevant problem in automated machine learning (AutoML) that allows finding suitable sequences of preprocessing and prediction models together with their optimal hyperparameters. This problem can be solved using evolutionary algorithms and, in particular, grammar-guided genetic programming (G3P). Current G3P approaches to AWC define a fixed grammar that formally specifies how workflow elements can be combined and which algorithms can be included. In this paper we present \ourmethod, an interactive G3P algorithm that allows users to dynamically modify the grammar to prune the search space and focus on their regions of interest. Our proposal is the first to combine the advantages of a G3P method with ideas from interactive optimisation and human-guided machine learning, an area little explored in the context of AutoML. To evaluate our approach, we present an experimental study in which 20 participants interact with \ourmethod to evolve workflows according to their preferences. Our results confirm that the collaboration between \ourmethod and humans allows us to find high-performance workflows in terms of accuracy that require less tuning time than those found without human intervention., Comment: 40 pages, 5 figures, 3 tables. Paper submitted to journal. Under review

We tackle the problem of bias mitigation of algorithmic decisions in a setting where both the output of the algorithm and the sensitive variable are continuous. Most of prior work deals with discrete sensitive variables, meaning that the biases are measured for subgroups of persons defined by a label, leaving out important algorithmic bias cases, where the sensitive variable is continuous. Typical examples are unfair decisions made with respect to the age or the financial status. In our work, we then propose a bias mitigation strategy for continuous sensitive variables, based on the notion of endogeneity which comes from the field of econometrics. In addition to solve this new problem, our bias mitigation strategy is a weakly supervised learning method which requires that a small portion of the data can be measured in a fair manner. It is model agnostic, in the sense that it does not make any hypothesis on the prediction model. It also makes use of a reasonably large amount of input observations and their corresponding predictions. Only a small fraction of the true output predictions should be known. This therefore limits the need for expert interventions. Results obtained on synthetic data show the effectiveness of our approach for examples as close as possible to real-life applications in econometrics., Comment: 30 pages, 25 figures

Consider a predictor, a learner, whose input is a stream of discrete items. The predictor's task, at every time point, is probabilistic multiclass prediction, i.e., to predict which item may occur next by outputting zero or more candidate items, each with a probability, after which the actual item is revealed and the predictor learns from this observation. To output probabilities, the predictor keeps track of the proportions of the items it has seen. The stream is unbounded and the predictor has finite limited space and we seek efficient prediction and update techniques: the set of items is unknown to the predictor and their totality can also grow unbounded. Moreover, there is non-stationarity: the underlying frequencies of items may change, substantially, from time to time. For instance, new items may start appearing and a few recently frequent items may cease to occur again. The predictor, being space-bounded, need only provide probabilities for those items with (currently) sufficiently high frequency, i.e., the salient items. This problem is motivated in the setting of prediction games, a self-supervised learning regime where concepts serve as both the predictors and the predictands, and the set of concepts grows over time, resulting in non-stationarities as new concepts are generated and used. We develop sparse multiclass moving average techniques designed to respond to such non-stationarities in a timely manner. One technique is based on the exponentiated moving average (EMA) and another is based on queuing a few count snapshots. We show that the combination, and in particular supporting dynamic predictand-specific learning rates, offers advantages in terms of faster change detection and convergence., Comment: 63 pages, 24 figures, 17 tables

Deceptive path planning (DPP) is the problem of designing a path that hides its true goal from an outside observer. Existing methods for DPP rely on unrealistic assumptions, such as global state observability and perfect model knowledge, and are typically problem-specific, meaning that even minor changes to a previously solved problem can force expensive computation of an entirely new solution. Given these drawbacks, such methods do not generalize to unseen problem instances, lack scalability to realistic problem sizes, and preclude both on-the-fly tunability of deception levels and real-time adaptivity to changing environments. In this paper, we propose a reinforcement learning (RL)-based scheme for training policies to perform DPP over arbitrary weighted graphs that overcomes these issues. The core of our approach is the introduction of a local perception model for the agent, a new state space representation distilling the key components of the DPP problem, the use of graph neural network-based policies to facilitate generalization and scaling, and the introduction of new deception bonuses that translate the deception objectives of classical methods to the RL setting. Through extensive experimentation we show that, without additional fine-tuning, at test time the resulting policies successfully generalize, scale, enjoy tunable levels of deception, and adapt in real-time to changes in the environment., Comment: 11 pages, 14 figures

Current approaches to prosthetic control are limited by their reliance on traditional methods, which lack real-time adaptability and intuitive responsiveness. These limitations are particularly pronounced in assistive technologies designed for individuals with diverse cognitive states and motor intentions. In this paper, we introduce a framework that leverages Reinforcement Learning (RL) with Deep Q-Networks (DQN) for classification tasks. Additionally, we present a preprocessing technique using the Common Spatial Pattern (CSP) for multiclass motor imagery (MI) classification in a One-Versus-The-Rest (OVR) manner. The subsequent 'csp space' transformation retains the temporal dimension of EEG signals, crucial for extracting discriminative features. The integration of DQN with a 1D-CNN-LSTM architecture optimizes the decision-making process in real-time, thereby enhancing the system's adaptability to the user's evolving needs and intentions. We elaborate on the data processing methods for two EEG motor imagery datasets. Our innovative model, RLEEGNet, incorporates a 1D-CNN-LSTM architecture as the Online Q-Network within the DQN, facilitating continuous adaptation and optimization of control strategies through feedback. This mechanism allows the system to learn optimal actions through trial and error, progressively improving its performance. RLEEGNet demonstrates high accuracy in classifying MI-EEG signals, achieving as high as 100% accuracy in MI tasks across both the GigaScience (3-class) and BCI-IV-2a (4-class) datasets. These results highlight the potential of combining DQN with a 1D-CNN-LSTM architecture to significantly enhance the adaptability and responsiveness of BCI systems., Comment: 23 pages, 1 figure, 6 tables

The process of extracting valuable and novel insights from raw data involves a series of complex steps. In the realm of Automated Machine Learning (AutoML), a significant research focus is on automating aspects of this process, specifically tasks like selecting algorithms and optimising their hyper-parameters. A particularly challenging task in AutoML is automatic workflow composition (AWC). AWC aims to identify the most effective sequence of data preprocessing and ML algorithms, coupled with their best hyper-parameters, for a specific dataset. However, existing AWC methods are limited in how many and in what ways they can combine algorithms within a workflow. Addressing this gap, this paper introduces EvoFlow, a grammar-based evolutionary approach for AWC. EvoFlow enhances the flexibility in designing workflow structures, empowering practitioners to select algorithms that best fit their specific requirements. EvoFlow stands out by integrating two innovative features. First, it employs a suite of genetic operators, designed specifically for AWC, to optimise both the structure of workflows and their hyper-parameters. Second, it implements a novel updating mechanism that enriches the variety of predictions made by different workflows. Promoting this diversity helps prevent the algorithm from overfitting. With this aim, EvoFlow builds an ensemble whose workflows differ in their misclassified instances. To evaluate EvoFlow's effectiveness, we carried out empirical validation using a set of classification benchmarks. We begin with an ablation study to demonstrate the enhanced performance attributable to EvoFlow's unique components. Then, we compare EvoFlow with other AWC approaches, encompassing both evolutionary and non-evolutionary techniques. Our findings show that EvoFlow's specialised genetic operators and updating mechanism substantially outperform current leading methods[..], Comment: 32 pages, 7 figures, 6 tables, journal paper

We seek to enable classic processing of continuous ultra-sparse spatiotemporal data generated by event-based sensors with dense machine learning models. We propose a novel hybrid pipeline composed of asynchronous sensing and synchronous processing that combines several ideas: (1) an embedding based on PointNet models -- the ALERT module -- that can continuously integrate new and dismiss old events thanks to a leakage mechanism, (2) a flexible readout of the embedded data that allows to feed any downstream model with always up-to-date features at any sampling rate, (3) exploiting the input sparsity in a patch-based approach inspired by Vision Transformer to optimize the efficiency of the method. These embeddings are then processed by a transformer model trained for object and gesture recognition. Using this approach, we achieve performances at the state-of-the-art with a lower latency than competitors. We also demonstrate that our asynchronous model can operate at any desired sampling rate., Comment: Originally published in the Proceedings of Machine Learning Research ICML 2024

Constructing the architecture of a neural network is a challenging pursuit for the machine learning community, and the dilemma of whether to go deeper or wider remains a persistent question. This paper explores a comparison between deeper neural networks (DeNNs) with a flexible number of layers and wider neural networks (WeNNs) with limited hidden layers, focusing on their optimal generalization error in Sobolev losses. Analytical investigations reveal that the architecture of a neural network can be significantly influenced by various factors, including the number of sample points, parameters within the neural networks, and the regularity of the loss function. Specifically, a higher number of parameters tends to favor WeNNs, while an increased number of sample points and greater regularity in the loss function lean towards the adoption of DeNNs. We ultimately apply this theory to address partial differential equations using deep Ritz and physics-informed neural network (PINN) methods, guiding the design of neural networks., Comment: arXiv admin note: text overlap with arXiv:2310.10766, arXiv:2305.08466

In traditional machine teaching, a teacher wants to teach a concept to a learner, by means of a finite set of examples, the witness set. But concepts can have many equivalent representations. This redundancy strongly affects the search space, to the extent that teacher and learner may not be able to easily determine the equivalence class of each representation. In this common situation, instead of teaching concepts, we explore the idea of teaching representations. We work with several teaching schemas that exploit representation and witness size (Eager, Greedy and Optimal) and analyze the gains in teaching effectiveness for some representational languages (DNF expressions and Turing-complete P3 programs). Our theoretical and experimental results indicate that there are various types of redundancy, handled better by the Greedy schema introduced here than by the Eager schema, although both can be arbitrarily far away from the Optimal. For P3 programs we found that witness sets are usually smaller than the programs they identify, which is an illuminating justification of why machine teaching from examples makes sense at all., Comment: 16 pages, 3 figures, 3 tables

Categorical random variables can faithfully represent the discrete and uncertain aspects of data as part of a discrete latent variable model. Learning in such models necessitates taking gradients with respect to the parameters of the categorical probability distributions, which is often intractable due to their combinatorial nature. A popular technique to estimate these otherwise intractable gradients is the Log-Derivative trick. This trick forms the basis of the well-known REINFORCE gradient estimator and its many extensions. While the Log-Derivative trick allows us to differentiate through samples drawn from categorical distributions, it does not take into account the discrete nature of the distribution itself. Our first contribution addresses this shortcoming by introducing the CatLog-Derivative trick - a variation of the Log-Derivative trick tailored towards categorical distributions. Secondly, we use the CatLog-Derivative trick to introduce IndeCateR, a novel and unbiased gradient estimator for the important case of products of independent categorical distributions with provably lower variance than REINFORCE. Thirdly, we empirically show that IndeCateR can be efficiently implemented and that its gradient estimates have significantly lower bias and variance for the same number of samples compared to the state of the art.

Battery degradation remains a pivotal concern in the energy storage domain, with machine learning emerging as a potent tool to drive forward insights and solutions. However, this intersection of electrochemical science and machine learning poses complex challenges. Machine learning experts often grapple with the intricacies of battery science, while battery researchers face hurdles in adapting intricate models tailored to specific datasets. Beyond this, a cohesive standard for battery degradation modeling, inclusive of data formats and evaluative benchmarks, is conspicuously absent. Recognizing these impediments, we present BatteryML - a one-step, all-encompass, and open-source platform designed to unify data preprocessing, feature extraction, and the implementation of both traditional and state-of-the-art models. This streamlined approach promises to enhance the practicality and efficiency of research applications. BatteryML seeks to fill this void, fostering an environment where experts from diverse specializations can collaboratively contribute, thus elevating the collective understanding and advancement of battery research.The code for our project is publicly available on GitHub at https://github.com/microsoft/BatteryML.

Regarding the rising number of people suffering from mental health illnesses in today's society, the importance of mental health cannot be overstated. Wearable sensors, which are increasingly widely available, provide a potential way to track and comprehend mental health issues. These gadgets not only monitor everyday activities but also continuously record vital signs like heart rate, perhaps providing information on a person's mental state. Recent research has used these sensors in conjunction with machine learning methods to identify patterns relating to different mental health conditions, highlighting the immense potential of this data beyond simple activity monitoring. In this research, we present a novel algorithm called the Hybrid Random forest - Neural network that has been tailored to evaluate sensor data from depressed patients. Our method has a noteworthy accuracy of 80\% when evaluated on a special dataset that included both unipolar and bipolar depressive patients as well as healthy controls. The findings highlight the algorithm's potential for reliably determining a person's depression condition using sensor data, making a substantial contribution to the area of mental health diagnostics., Comment: 8 pages

Threat hunting is a proactive methodology for exploring, detecting and mitigating cyberattacks within complex environments. As opposed to conventional detection systems, threat hunting strategies assume adversaries have infiltrated the system; as a result they proactively search out any unusual patterns or activities which might indicate intrusion attempts. Historically, this endeavour has been pursued using three investigation methodologies: (1) Hypothesis-Driven Investigations; (2) Indicator of Compromise (IOC); and (3) High-level machine learning analysis-based approaches. Therefore, this paper introduces a novel machine learning paradigm known as Threat Trekker. This proposal utilizes connectors to feed data directly into an event streaming channel for processing by the algorithm and provide feedback back into its host network. Conclusions drawn from these experiments clearly establish the efficacy of employing machine learning for classifying more subtle attacks., Comment: I am disseminating this outcome to all of you, despite the fact that the results may appear somewhat idealistic, given that certain datasets utilized for the training of the machine learning model comprise simulated data

Recent research efforts on semantic communication have mostly considered accuracy as a main problem for optimizing goal-oriented communication systems. However, these approaches introduce a paradox: the accuracy of artificial intelligence (AI) tasks should naturally emerge through training rather than being dictated by network constraints. Acknowledging this dilemma, this work introduces an innovative approach that leverages the rate-distortion theory to analyze distortions induced by communication and semantic compression, thereby analyzing the learning process. Specifically, we examine the distribution shift between the original data and the distorted data, thus assessing its impact on the AI model's performance. Founding upon this analysis, we can preemptively estimate the empirical accuracy of AI tasks, making the goal-oriented semantic communication problem feasible. To achieve this objective, we present the theoretical foundation of our approach, accompanied by simulations and experiments that demonstrate its effectiveness. The experimental results indicate that our proposed method enables accurate AI task performance while adhering to network constraints, establishing it as a valuable contribution to the field of signal processing. Furthermore, this work advances research in goal-oriented semantic communication and highlights the significance of data-driven approaches in optimizing the performance of intelligent systems., Comment: 15 pages; 11 figures, 2 tables

In this paper, we present Layer-wise Feedback Propagation (LFP), a novel training approach for neural-network-like predictors that utilizes explainability, specifically Layer-wise Relevance Propagation(LRP), to assign rewards to individual connections based on their respective contributions to solving a given task. This differs from traditional gradient descent, which updates parameters towards anestimated loss minimum. LFP distributes a reward signal throughout the model without the need for gradient computations. It then strengthens structures that receive positive feedback while reducingthe influence of structures that receive negative feedback. We establish the convergence of LFP theoretically and empirically, and demonstrate its effectiveness in achieving comparable performance to gradient descent on various models and datasets. Notably, LFP overcomes certain limitations associated with gradient-based methods, such as reliance on meaningful derivatives. We further investigate how the different LRP-rules can be extended to LFP, what their effects are on training, as well as potential applications, such as training models with no meaningful derivatives, e.g., step-function activated Spiking Neural Networks (SNNs), or for transfer learning, to efficiently utilize existing knowledge.

Intrusion detection systems (IDSs) are essential elements of IT systems. Their key component is a classification module that continuously evaluates some features of the network traffic and identifies possible threats. Its efficiency is greatly affected by the right selection of the features to be monitored. Therefore, the identification of a minimal set of features that are necessary to safely distinguish malicious traffic from benign traffic is indispensable in the course of the development of an IDS. This paper presents the preprocessing and feature selection workflow as well as its results in the case of the CSE-CIC-IDS2018 on AWS dataset, focusing on five attack types. To identify the relevant features, six feature selection methods were applied, and the final ranking of the features was elaborated based on their average score. Next, several subsets of the features were formed based on different ranking threshold values, and each subset was tried with five classification algorithms to determine the optimal feature set for each attack type. During the evaluation, four widely used metrics were taken into consideration., Comment: 24 pages

In this paper we revisit the fixed-confidence identification of the Pareto optimal set in a multi-objective multi-armed bandit model. As the sample complexity to identify the exact Pareto set can be very large, a relaxation allowing to output some additional near-optimal arms has been studied. In this work we also tackle alternative relaxations that allow instead to identify a relevant subset of the Pareto set. Notably, we propose a single sampling strategy, called Adaptive Pareto Exploration, that can be used in conjunction with different stopping rules to take into account different relaxations of the Pareto Set Identification problem. We analyze the sample complexity of these different combinations, quantifying in particular the reduction in sample complexity that occurs when one seeks to identify at most $k$ Pareto optimal arms. We showcase the good practical performance of Adaptive Pareto Exploration on a real-world scenario, in which we adaptively explore several vaccination strategies against Covid-19 in order to find the optimal ones when multiple immunogenicity criteria are taken into account.