Consensus-based optimization (CBO) has established itself as an efficient gradient-free optimization scheme, with attractive mathematical properties, such as mean-field convergence results for non-convex loss functions. In this work, we study CBO in the context of closed-box adversarial attacks, which are imperceptible input perturbations that aim to fool a classifier, without accessing its gradient. Our contribution is to establish a connection between the so-called consensus hopping as introduced by Riedl et al. and natural evolution strategies (NES) commonly applied in the context of adversarial attacks and to rigorously relate both methods to gradient-based optimization schemes. Beyond that, we provide a comprehensive experimental study that shows that despite the conceptual similarities, CBO can outperform NES and other evolutionary strategies in certain scenarios.

In the face of evolving cyber threats such as malware, ransomware and phishing, autonomous cybersecurity defense (ACD) systems have become essential for real-time threat detection and response with optional human intervention. However, existing ACD systems rely on limiting assumptions, particularly the stationarity of the underlying network dynamics. In real-world scenarios, network topologies can change due to actions taken by attackers or defenders, system failures, or time evolution of networks, leading to failures in the adaptive capabilities of current defense agents. Moreover, many agents are trained on static environments, resulting in overfitting to specific topologies, which hampers their ability to generalize to out-of-distribution network topologies. This work addresses these challenges by exploring methods for developing agents to learn generalizable policies across dynamic network environments -- general ACD (GACD).

The rise of API-only access to state-of-the-art LLMs highlights the need for effective black-box jailbreak methods to identify model vulnerabilities in real-world settings. Without a principled objective for gradient-based optimization, most existing approaches rely on genetic algorithms, which are limited by their initialization and dependence on manually curated prompt pools. Furthermore, these methods require individual optimization for each prompt, failing to provide a comprehensive characterization of model vulnerabilities. To address this gap, we introduce VERA: Variational infErence fRamework for jAilbreaking. VERA casts black-box jailbreak prompting as a variational inference problem, training a small attacker LLM to approximate the target LLM's posterior over adversarial prompts. Once trained, the attacker can generate diverse, fluent jailbreak prompts for a target query without re-optimization. Experimental results show that VERA achieves strong performance across a range of target LLMs, highlighting the value of probabilistic inference for adversarial prompt generation.

We present `GL-LowPopArt`, a novel Catoni-style estimator for generalized low-rank trace regression. Building on `LowPopArt` (Jang et al., 2024), it employs a two-stage approach: nuclear norm regularization followed by matrix Catoni estimation. We establish state-of-the-art estimation error bounds, surpassing existing guarantees (Fan et al., 2019; Kang et al., 2022), and reveal a novel experimental design objective, $\mathrm{GL}(π)$. The key technical challenge is controlling bias from the nonlinear inverse link function, which we address by our two-stage approach. We prove a *local* minimax lower bound, showing that our `GL-LowPopArt` enjoys instance-wise optimality up to the condition number of the ground-truth Hessian. Applications include generalized linear matrix completion, where `GL-LowPopArt` achieves a state-of-the-art Frobenius error guarantee, and **bilinear dueling bandits**, a novel setting inspired by general preference learning (Zhang et al., 2024). Our analysis of a `GL-LowPopArt`-based explore-then-commit algorithm reveals a new, potentially interesting problem-dependent quantity, along with improved Borda regret bound than vectorization (Wu et al., 2024).

The Laser Interferometer Space Antenna (LISA) is expected to detect thousands of individually resolved gravitational wave sources, overlapping in time and frequency, on top of unresolved astrophysical and/or primordial backgrounds. Disentangling resolved sources from backgrounds and extracting their parameters in a computationally intensive "global fit" is normally regarded as a necessary step toward reconstructing the properties of the underlying astrophysical populations. Here, we show that it is possible to infer the properties of the most numerous population of LISA sources - Galactic double white dwarfs - directly from the frequency (or, equivalently, time) strain series, by using a simulation-based approach that bypasses the global fit entirely. By training a normalizing flow on a custom-designed compression of simulated LISA frequency series from the Galactic double white dwarf population, we demonstrate how to infer the posterior distribution of population parameters (e.g., mass function, frequency, and spatial distributions). This allows for extracting information on the population parameters from both resolved and unresolved sources simultaneously and in a computationally efficient manner. Our approach to target population properties directly can be readily extended to other source classes (e.g., massive and stellar-mass black holes, extreme mass ratio inspirals), provided fast simulations are available, and to scenarios involving non-Gaussian or non-stationary noise (e.g., data gaps).

Scientific information expresses human understanding of nature. This knowledge is largely disseminated in different forms of text, including scientific papers, news articles, and discourse among people on social media. While important for accelerating our pursuit of knowledge, not all scientific text is faithful to the underlying science. As the volume of this text has burgeoned online in recent years, it has become a problem of societal importance to be able to identify the faithfulness of a given piece of scientific text automatically. This thesis is concerned with the cultivation of datasets, methods, and tools for machine understanding of scientific language, in order to analyze and understand science communication at scale. To arrive at this, I present several contributions in three areas of natural language processing and machine learning: automatic fact checking, learning with limited data, and scientific text processing. These contributions include new methods and resources for identifying check-worthy claims, adversarial claim generation, multi-source domain adaptation, learning from crowd-sourced labels, cite-worthiness detection, zero-shot scientific fact checking, detecting exaggerated scientific claims, and modeling degrees of information change in science communication. Critically, I demonstrate how the research outputs of this thesis are useful for effectively learning from limited amounts of scientific text in order to identify misinformative scientific statements and generate new insights into the science communication process

Autonomous Driving Systems (ADSs) continue to face safety-critical risks due to the inherent limitations in their design and performance capabilities. Online repair plays a crucial role in mitigating such limitations, ensuring the runtime safety and reliability of ADSs. Existing online repair solutions enforce ADS compliance by transforming unacceptable trajectories into acceptable ones based on predefined specifications, such as rule-based constraints or training datasets. However, these approaches often lack generalizability, adaptability and tend to be overly conservative, resulting in ineffective repairs that not only fail to mitigate safety risks sufficiently but also degrade the overall driving experience. To address this issue, we propose Adaptive Decision Repair (ADReFT), a novel and effective repair method that identifies safety-critical states through offline learning from failed tests and generates appropriate mitigation actions to improve ADS safety. Specifically, ADReFT incorporates a transformer-based model with two joint heads, State Monitor and Decision Adapter, designed to capture complex driving environment interactions to evaluate state safety severity and generate adaptive repair actions. Given the absence of oracles for state safety identification, we first pretrain ADReFT using supervised learning with coarse annotations, i.e., labeling states preceding violations as positive samples and others as negative samples. It establishes ADReFT's foundational capability to mitigate safety-critical violations, though it may result in somewhat conservative mitigation strategies. Therefore, we subsequently finetune ADReFT using reinforcement learning to improve its initial capability and generate more precise and contextually appropriate repair decisions. Our evaluation results illustrate that ADReFT achieves better repair performance.

Robot end users increasingly require accessible means of specifying tasks for robots to perform. Two common end-user programming paradigms include drag-and-drop interfaces and natural language programming. Although natural language interfaces harness an intuitive form of human communication, drag-and-drop interfaces enable users to meticulously and precisely dictate the key actions of the robot's task. In this paper, we investigate the degree to which both approaches can be combined. Specifically, we construct a large language model (LLM)-based pipeline that accepts natural language as input and produces human-like action sequences as output, specified at a level of granularity that a human would produce. We then compare these generated action sequences to another dataset of hand-specified action sequences. Although our results reveal that larger models tend to outperform smaller ones in the production of human-like action sequences, smaller models nonetheless achieve satisfactory performance.

Recent advances in machine learning, particularly the emergence of foundation models, are leading to new opportunities to develop technology-based solutions to societal problems. However, the reasoning and inner workings of today's complex AI models are not transparent to the user, and there are no safety guarantees regarding their predictions. Consequently, to fulfill the promise of AI, we must address the following scientific challenge: how to develop AI-based systems that are not only accurate and performant but also safe and trustworthy? The criticality of safe operation is particularly evident for autonomous systems for control and robotics, and was the catalyst for the Safe Learning Enabled Systems (SLES) program at NSF. For the broader class of AI applications, such as users interacting with chatbots and clinicians receiving treatment recommendations, safety is, while no less important, less well-defined with context-dependent interpretations. This motivated the organization of a day-long workshop, held at University of Pennsylvania on February 26, 2025, to bring together investigators funded by the NSF SLES program with a broader pool of researchers studying AI safety. This report is the result of the discussions in the working groups that addressed different aspects of safety at the workshop. The report articulates a new research agenda focused on developing theory, methods, and tools that will provide the foundations of the next generation of AI-enabled systems.

The integration of the Internet of Everything (IoX) and Artificial General Intelligence (AGI) has given rise to a transformative paradigm aimed at addressing critical bottlenecks across sensing, network, and application layers in Cyber-Physical-Social Thinking (CPST) ecosystems. In this survey, we provide a systematic and comprehensive review of AGI-enhanced IoX research, focusing on three key components: sensing-layer data management, network-layer protocol optimization, and application-layer decision-making frameworks. Specifically, this survey explores how AGI can mitigate IoX bottlenecks challenges by leveraging adaptive sensor fusion, edge preprocessing, and selective attention mechanisms at the sensing layer, while resolving network-layer issues such as protocol heterogeneity and dynamic spectrum management, neuro-symbolic reasoning, active inference, and causal reasoning, Furthermore, the survey examines AGI-enabled frameworks for managing identity and relationship explosion. Key findings suggest that AGI-driven strategies, such as adaptive sensor fusion, edge preprocessing, and semantic modeling, offer novel solutions to sensing-layer data overload, network-layer protocol heterogeneity, and application-layer identity explosion. The survey underscores the importance of cross-layer integration, quantum-enabled communication, and ethical governance frameworks for future AGI-enabled IoX systems. Finally, the survey identifies unresolved challenges, such as computational requirements, scalability, and real-world validation, calling for further research to fully realize AGI's potential in addressing IoX bottlenecks. we believe AGI-enhanced IoX is emerging as a critical research field at the intersection of interconnected systems and advanced AI.