With the widespread deployment of Multimodal Large Language Models (MLLMs) for visual-reasoning tasks, improving their safety has become crucial. Recent research indicates that despite training-time safety alignment, these models remain vulnerable to jailbreak attacks. In this work, we first highlight an important safety gap to describe that alignment achieved solely through safety training may be insufficient against jailbreak attacks. To address this vulnerability, we propose Immune, an inference-time defense framework that leverages a safe reward model through controlled decoding to defend against jailbreak attacks. Additionally, we provide a mathematical characterization of Immune, offering provable guarantees against jailbreaks. Extensive evaluations on diverse jailbreak benchmarks using recent MLLMs reveal that Immune effectively enhances model safety while preserving the model's original capabilities. For instance, against text-based jailbreak attacks on LLaVA-1.6, Immune reduces the attack success rate by 57.82% and 16.78% compared to the base MLLM and state-of-the-art defense strategy, respectively.

Background: Despite the current ubiquity of Large Language Models (LLMs) across the medical domain, there is a surprising lack of studies which address their reasoning behaviour. We emphasise the importance of understanding reasoning behaviour as opposed to high-level prediction accuracies, since it is equivalent to explainable AI (XAI) in this context. In particular, achieving XAI in medical LLMs used in the clinical domain will have a significant impact across the healthcare sector. Results: Therefore, we define the concept of reasoning behaviour in the specific context of medical LLMs. We then categorise and discuss the current state of the art of methods which evaluate reasoning behaviour in medical LLMs. Finally, we propose theoretical frameworks which can empower medical professionals or machine learning engineers to gain insight into the low-level reasoning operations of these previously obscure models. Conclusion: The subsequent increased transparency and trust in medical machine learning models by clinicians as well as patients will accelerate the integration, application as well as further development of medical AI for the healthcare system as a whole

We review the recent literature (January 2022- October 2024) in South Asian languages on text-based language processing, multimodal models, and speech processing, and provide a spotlight analysis focused on 21 low-resource South Asian languages, namely Saraiki, Assamese, Balochi, Bhojpuri, Bodo, Burmese, Chhattisgarhi, Dhivehi, Gujarati, Kannada, Kashmiri, Konkani, Khasi, Malayalam, Meitei, Nepali, Odia, Pashto, Rajasthani, Sindhi, and Telugu. We identify trends, challenges, and future research directions, using a step-wise approach that incorporates relevance classification and clustering based on large language models (LLMs). Our goal is to provide a breadth-first overview of the recent developments in South Asian language technologies to NLP researchers interested in working with South Asian languages.

This systematic literature review investigates perceptions, concerns, and expectations of young digital citizens regarding privacy in artificial intelligence (AI) systems, focusing on social media platforms, educational technology, gaming systems, and recommendation algorithms. Using a rigorous methodology, the review started with 2,000 papers, narrowed down to 552 after initial screening, and finally refined to 108 for detailed analysis. Data extraction focused on privacy concerns, data-sharing practices, the balance between privacy and utility, trust factors in AI, transparency expectations, and strategies to enhance user control over personal data. Findings reveal significant privacy concerns among young users, including a perceived lack of control over personal information, potential misuse of data by AI, and fears of data breaches and unauthorized access. These issues are worsened by unclear data collection practices and insufficient transparency in AI applications. The intention to share data is closely associated with perceived benefits and data protection assurances. The study also highlights the role of parental mediation and the need for comprehensive education on data privacy. Balancing privacy and utility in AI applications is crucial, as young digital citizens value personalized services but remain wary of privacy risks. Trust in AI is significantly influenced by transparency, reliability, predictable behavior, and clear communication about data usage. Strategies to improve user control over personal data include access to and correction of data, clear consent mechanisms, and robust data protection assurances. The review identifies research gaps and suggests future directions, such as longitudinal studies, multicultural comparisons, and the development of ethical AI frameworks.

With the rise of foundation models, there is growing concern about their potential social impacts. Social science has a long history of studying the social impacts of transformative technologies in terms of pre-existing systems of power and how these systems are disrupted or reinforced by new technologies. In this position paper, we build on prior work studying the social impacts of earlier technologies to propose a conceptual framework studying foundation models as sociotechnical systems, incorporating social science expertise to better understand how these models affect systems of power, anticipate the impacts of deploying these models in various applications, and study the effectiveness of technical interventions intended to mitigate social harms. We advocate for an interdisciplinary and collaborative research paradigm between AI and social science across all stages of foundation model research and development to promote socially responsible research practices and use cases, and outline several strategies to facilitate such research.

There is a significant body of work looking at the ethical considerations of large language models (LLMs): critiquing tools to measure performance and harms; proposing toolkits to aid in ideation; discussing the risks to workers; considering legislation around privacy and security etc. As yet there is no work that integrates these resources into a single practical guide that focuses on LLMs; we attempt this ambitious goal. We introduce 'LLM Ethics Whitepaper', which we provide as an open and living resource for NLP practitioners, and those tasked with evaluating the ethical implications of others' work. Our goal is to translate ethics literature into concrete recommendations and provocations for thinking with clear first steps, aimed at computer scientists. 'LLM Ethics Whitepaper' distils a thorough literature review into clear Do's and Don'ts, which we present also in this paper. We likewise identify useful toolkits to support ethical work. We refer the interested reader to the full LLM Ethics Whitepaper, which provides a succinct discussion of ethical considerations at each stage in a project lifecycle, as well as citations for the hundreds of papers from which we drew our recommendations. The present paper can be thought of as a pocket guide to conducting ethical research with LLMs.

Natural Language Processing (NLP) is widely used in fields like machine translation and sentiment analysis. However, traditional NLP models struggle with accuracy and efficiency. This paper introduces Deep Convolutional Neural Networks (DCNN) into NLP to address these issues. By integrating DCNN, machine learning (ML) algorithms, and generative adversarial networks (GAN), the study improves language understanding, reduces ambiguity, and enhances task performance. The high-performance NLP model shows a 10% improvement in segmentation accuracy and a 4% increase in recall rate compared to traditional models. This integrated approach excels in tasks such as word segmentation, part-of-speech tagging, machine translation, and text classification, offering better recognition accuracy and processing efficiency.

The growing threats of uncertainties, anomalies, and cyberattacks on power grids are driving a critical need to advance situational awareness which allows system operators to form a complete and accurate picture of the present and future state. Simulation and estimation are foundational tools in this process. However, existing tools lack the robustness and efficiency required to achieve the level of situational awareness needed for the ever-evolving threat landscape. Industry-standard (steady-state) simulators are not robust to blackouts, often leading to non-converging or non-actionable results. Estimation tools lack robustness to anomalous data, returning erroneous system states. Efficiency is the other major concern as nonlinearities and scalability issues make large systems slow to converge. This thesis addresses robustness and efficiency gaps through a dual-fold contribution. We first address the inherent limitations in the existing physics-based and data-driven worlds; and then transcend the boundaries of conventional algorithmic design in the direction of a new paradigm -- Physics-ML Synergy -- which integrates the strengths of the two worlds. Our approaches are built on circuit formulation which provides a unified framework that applies to both transmission and distribution. Sparse optimization acts as the key enabler to make these tools intrinsically robust and immune to random threats, pinpointing dominant sources of (random) blackouts and data errors. Further, we explore sparsity-exploiting optimizations to develop lightweight ML models whose prediction and detection capabilities are a complement to physics-based tools; and whose lightweight designs advance generalization and scalability. Finally, Physics-ML Synergy brings robustness and efficiency further against targeted cyberthreats, by interconnecting our physics-based tools with lightweight ML.

In multi-agent environments, agents often struggle to learn optimal policies due to sparse or delayed global rewards, particularly in long-horizon tasks where it is challenging to evaluate actions at intermediate time steps. We introduce Temporal-Agent Reward Redistribution (TAR$^2$), a novel approach designed to address the agent-temporal credit assignment problem by redistributing sparse rewards both temporally and across agents. TAR$^2$ decomposes sparse global rewards into time-step-specific rewards and calculates agent-specific contributions to these rewards. We theoretically prove that TAR$^2$ is equivalent to potential-based reward shaping, ensuring that the optimal policy remains unchanged. Empirical results demonstrate that TAR$^2$ stabilizes and accelerates the learning process. Additionally, we show that when TAR$^2$ is integrated with single-agent reinforcement learning algorithms, it performs as well as or better than traditional multi-agent reinforcement learning methods.

The optimization of large-scale multibody systems is a numerically challenging task, in particular when considering multiple conflicting criteria at the same time. In this situation, we need to approximate the Pareto set of optimal compromises, which is significantly more expensive than finding a single optimum in single-objective optimization. To prevent large costs, the usage of surrogate models, constructed from a small but informative number of expensive model evaluations, is a very popular and widely studied approach. The central challenge then is to ensure a high quality (that is, near-optimality) of the solutions that were obtained using the surrogate model, which can be hard to guarantee with a single pre-computed surrogate. We present a back-and-forth approach between surrogate modeling and multi-objective optimization to improve the quality of the obtained solutions. Using the example of an expensive-to-evaluate multibody system, we compare different strategies regarding multi-objective optimization, sampling and also surrogate modeling, to identify the most promising approach in terms of computational efficiency and solution quality.