Coordination among connected and autonomous vehicles (CAVs) is advancing due to developments in control and communication technologies. However, much of the current work is based on oversimplified and unrealistic task-specific assumptions, which may introduce vulnerabilities. This is critical because CAVs not only interact with their environment but are also integral parts of it. Insufficient exploration can result in policies that carry latent risks, highlighting the need for methods that explore the environment both extensively and efficiently. This work introduces OPTIMA, a novel distributed reinforcement learning framework for cooperative autonomous vehicle tasks. OPTIMA alternates between thorough data sampling from environmental interactions and multi-agent reinforcement learning algorithms to optimize CAV cooperation, emphasizing both safety and efficiency. Our goal is to improve the generality and performance of CAVs in highly complex and crowded scenarios. Furthermore, the industrial-scale distributed training system easily adapts to different algorithms, reward functions, and strategies.

Human trajectory anomaly detection has become increasingly important across a wide range of applications, including security surveillance and public health. However, existing trajectory anomaly detection methods are primarily focused on vehicle-level traffic, while human-level trajectory anomaly detection remains under-explored. Since human trajectory data is often very sparse, machine learning methods have become the preferred approach for identifying complex patterns. However, concerns regarding potential biases and the robustness of these models have intensified the demand for more transparent and explainable alternatives. In response to these challenges, our research focuses on developing a lightweight anomaly detection model specifically designed to detect anomalies in human trajectories. We propose a Neural Collaborative Filtering approach to model and predict normal mobility. Our method is designed to model users' daily patterns of life without requiring prior knowledge, thereby enhancing performance in scenarios where data is sparse or incomplete, such as in cold start situations. Our algorithm consists of two main modules. The first is the collaborative filtering module, which applies collaborative filtering to model normal mobility of individual humans to places of interest. The second is the neural module, responsible for interpreting the complex spatio-temporal relationships inherent in human trajectory data. To validate our approach, we conducted extensive experiments using simulated and real-world datasets comparing to numerous state-of-the-art trajectory anomaly detection approaches.

Safe reinforcement learning (RL) is a promising approach for many real-world decision-making problems where ensuring safety is a critical necessity. In safe RL research, while expected cumulative safety constraints (ECSCs) are typically the first choices, chance constraints are often more pragmatic for incorporating safety under uncertainties. This paper proposes a \textit{flipping-based policy} for Chance-Constrained Markov Decision Processes (CCMDPs). The flipping-based policy selects the next action by tossing a potentially distorted coin between two action candidates. The probability of the flip and the two action candidates vary depending on the state. We establish a Bellman equation for CCMDPs and further prove the existence of a flipping-based policy within the optimal solution sets. Since solving the problem with joint chance constraints is challenging in practice, we then prove that joint chance constraints can be approximated into Expected Cumulative Safety Constraints (ECSCs) and that there exists a flipping-based policy in the optimal solution sets for constrained MDPs with ECSCs. As a specific instance of practical implementations, we present a framework for adapting constrained policy optimization to train a flipping-based policy. This framework can be applied to other safe RL algorithms. We demonstrate that the flipping-based policy can improve the performance of the existing safe RL algorithms under the same limits of safety constraints on Safety Gym benchmarks.

Lithium-ion batteries are a key energy storage technology driving revolutions in mobile electronics, electric vehicles and renewable energy storage. Capacity retention is a vital performance measure that is frequently utilized to assess whether these batteries have approached their end-of-life. Machine learning (ML) offers a powerful tool for predicting capacity degradation based on past data, and, potentially, prior physical knowledge, but the degree to which an ML prediction can be trusted is of significant practical importance in situations where consequential decisions must be made based on battery state of health. This study explores the efficacy of fully Bayesian machine learning in forecasting battery health with the quantification of uncertainty in its predictions. Specifically, we implemented three probabilistic ML approaches and evaluated the accuracy of their predictions and uncertainty estimates: a standard Gaussian process (GP), a structured Gaussian process (sGP), and a fully Bayesian neural network (BNN). In typical applications of GP and sGP, their hyperparameters are learned from a single sample while, in contrast, BNNs are typically pre-trained on an existing dataset to learn the weight distributions before being used for inference. This difference in methodology gives the BNN an advantage in learning global trends in a dataset and makes BNNs a good choice when training data is available. However, we show that pre-training can also be leveraged for GP and sGP approaches to learn the prior distributions of the hyperparameters and that in the case of the pre-trained sGP, similar accuracy and improved uncertainty estimation compared to the BNN can be achieved. This approach offers a framework for a broad range of probabilistic machine learning scenarios where past data is available and can be used to learn priors for (hyper)parameters of probabilistic ML models.

In Computer science and across various engineering fields, noise is often considered a nuisance and annoyance. It distorts details and makes data less accurate. In the past, the goal has often been to eliminate noise with the goal to make systems more reliable and accurate. But views on noise are changing. New findings suggest that noise can actually enhance and advance technologies in many areas, making us see it not just as a disruption but as a way to improve system performance. Thus, once unwanted and hard to control, noise now appears to be a key player in improving the performance of complex information processing systems [22]. This phenomena is often known as Stochastic Resonance, which helps clear up signals, improve image quality, and strengthen models in machine learning [7, 22, 101]. This duality of noise -- both a problem and a benefit -- highlights the tricky role of noise while optimizing advanced neural networks and machine learning models.

This study evaluates the effectiveness of machine learning (ML) and deep learning (DL) models in detecting COVID-19-related misinformation on online social networks (OSNs), aiming to develop more effective tools for countering the spread of health misinformation during the pan-demic. The study trained and tested various ML classifiers (Naive Bayes, SVM, Random Forest, etc.), DL models (CNN, LSTM, hybrid CNN+LSTM), and pretrained language models (DistilBERT, RoBERTa) on the "COVID19-FNIR DATASET". These models were evaluated for accuracy, F1 score, recall, precision, and ROC, and used preprocessing techniques like stemming and lemmatization. The results showed SVM performed well, achieving a 94.41% F1-score. DL models with Word2Vec embeddings exceeded 98% in all performance metrics (accuracy, F1 score, recall, precision & ROC). The CNN+LSTM hybrid models also exceeded 98% across performance metrics, outperforming pretrained models like DistilBERT and RoBERTa. Our study concludes that DL and hybrid DL models are more effective than conventional ML algorithms for detecting COVID-19 misinformation on OSNs. The findings highlight the importance of advanced neural network approaches and large-scale pretraining in misinformation detection. Future research should optimize these models for various misinformation types and adapt to changing OSNs, aiding in combating health misinformation.

In this work, we tackle a challenging and extreme form of subpopulation shift, which is termed compositional shift. Under compositional shifts, some combinations of attributes are totally absent from the training distribution but present in the test distribution. We model the data with flexible additive energy distributions, where each energy term represents an attribute, and derive a simple alternative to empirical risk minimization termed compositional risk minimization (CRM). We first train an additive energy classifier to predict the multiple attributes and then adjust this classifier to tackle compositional shifts. We provide an extensive theoretical analysis of CRM, where we show that our proposal extrapolates to special affine hulls of seen attribute combinations. Empirical evaluations on benchmark datasets confirms the improved robustness of CRM compared to other methods from the literature designed to tackle various forms of subpopulation shifts.

Reinforcement Learning from Human Feedback (RLHF) has emerged as a pivotal tool for aligning large language models (LLMs) with human preferences. Direct Preference Optimization (DPO), one of the most popular approaches, formulates RLHF as a policy optimization problem without explicitly estimating the reward function. It overcomes the stability and efficiency issues of two-step approaches, which typically involve first estimating the reward function and then optimizing the policy via proximal policy optimization (PPO). Since RLHF is essentially an optimization problem, and it is well-known that momentum techniques can accelerate optimization both theoretically and empirically, a natural question arises: Can RLHF be accelerated by momentum? This paper answers this question in the affirmative. In detail, we first show that the iterative preference optimization method can be viewed as a proximal point method. Based on this observation, we propose a general Accelerated Preference Optimization (APO) framework, which unifies many existing preference optimization algorithms and employs Nesterov's momentum technique to speed up the alignment of LLMs. Theoretically, we demonstrate that APO can achieve a faster convergence rate than the standard iterative preference optimization methods, including DPO and Self-Play Preference Optimization (SPPO). Empirically, we show the superiority of APO over DPO, iterative DPO, and other strong baselines for RLHF on the AlpacaEval 2.0 benchmark.

In reinforcement learning, if the agent's reward differs from the designers' true utility, even only rarely, the state distribution resulting from the agent's policy can be very bad, in theory and in practice. When RL policies would devolve into undesired behavior, a common countermeasure is KL regularization to a trusted policy ("Don't do anything I wouldn't do"). All current cutting-edge language models are RL agents that are KL-regularized to a "base policy" that is purely predictive. Unfortunately, we demonstrate that when this base policy is a Bayesian predictive model of a trusted policy, the KL constraint is no longer reliable for controlling the behavior of an advanced RL agent. We demonstrate this theoretically using algorithmic information theory, and while systems today are too weak to exhibit this theorized failure precisely, we RL-finetune a language model and find evidence that our formal results are plausibly relevant in practice. We also propose a theoretical alternative that avoids this problem by replacing the "Don't do anything I wouldn't do" principle with "Don't do anything I mightn't do".

With the growing importance of customer service in contemporary business, recognizing the intents behind service dialogues has become essential for the strategic success of enterprises. However, the nature of dialogue data varies significantly across different scenarios, and implementing an intent parser for a specific domain often involves tedious feature engineering and a heavy workload of data labeling. In this paper, we propose a novel Neural-Bayesian Program Learning model named Dialogue-Intent Parser (DI-Parser), which specializes in intent parsing under data-hungry settings and offers promising performance improvements. DI-Parser effectively utilizes data from multiple sources in a "Learning to Learn" manner and harnesses the "wisdom of the crowd" through few-shot learning capabilities on human-annotated datasets. Experimental results demonstrate that DI-Parser outperforms state-of-the-art deep learning models and offers practical advantages for industrial-scale applications.