Recent advancements in Large Language Models (LLMs) have led to their widespread adoption in daily applications. Despite their impressive capabilities, they remain vulnerable to adversarial attacks, as even minor meaning-preserving changes such as synonym substitutions can lead to incorrect predictions. As a result, certifying the robustness of LLMs against such adversarial prompts is of vital importance. Existing approaches focused on word deletion or simple denoising strategies to achieve robustness certification. However, these methods face two critical limitations: (1) they yield loose robustness bounds due to the lack of semantic validation for perturbed outputs and (2) they suffer from high computational costs due to repeated sampling. To address these limitations, we propose CluCERT, a novel framework for certifying LLM robustness via clustering-guided denoising smoothing. Specifically, to achieve tighter certified bounds, we introduce a semantic clustering filter that reduces noisy samples and retains meaningful perturbations, supported by theoretical analysis. Furthermore, we enhance computational efficiency through two mechanisms: a refine module that extracts core semantics, and a fast synonym substitution strategy that accelerates the denoising process. Finally, we conduct extensive experiments on various downstream tasks and jailbreak defense scenarios. Experimental results demonstrate that our method outperforms existing certified approaches in both robustness bounds and computational efficiency.

Language Models (LMs) struggle with complex, interdependent instructions, particularly in high-stakes domains like finance where precision is critical. We introduce FIFE, a novel, high-difficulty benchmark designed to assess LM instruction-following capabilities for financial analysis tasks. FIFE comprises 88 human-authored prompts and employs a verification system with chainable, verifiable constraints for fine-grained reward signals. We evaluate 53 models (proprietary, open-weight, open-source) in a zero-shot setting. Our key findings reveal a clear performance hierarchy: the top open-weight model (76.1 strict / 79.5 loose) surpasses the leading proprietary system (65.9 strict / 70.5 loose), while the best open-source models lag significantly (45.5 strict / 48.9 loose). However, even top-performing models struggle with FIFE's complex requirements, failing to achieve perfect compliance. We release our dataset and code as an open-source resource to promote research in Reinforcement Learning for the financial domain.

Low-Rank Adaptation (LoRA) enables efficient Continual Learning but often suffers from catastrophic forgetting due to destructive interference between tasks. Our analysis reveals that this degradation is primarily driven by antagonistic directional updates where new task gradients directly oppose the historical weight trajectory. To address this, we propose PS-LoRA (Parameter Stability LoRA), a framework designed to resolve conflicts by aligning updates within the optimization subspace. Our approach employs a dual-regularization objective that penalizes conflicting directions and constrains magnitude deviations to ensure consistency with prior knowledge. Additionally, we implement a magnitude-based merging strategy to consolidate sequential adapters into a robust representation without retraining. Experiments on NLP and Vision benchmarks show that PS-LoRA outperforms state-of-the-art methods by preserving the stability of learned representations while efficiently adapting to new domains.

We introduce a unified benchmarking framework focused on evaluating EEG-based foundation models in clinical applications. The benchmark spans 11 well-defined diagnostic tasks across 14 publicly available EEG datasets, including epilepsy, schizophrenia, Parkinson's disease, OCD, and mild traumatic brain injury. It features minimal preprocessing, standardized evaluation protocols, and enables side-by-side comparisons of classical baselines and modern foundation models. Our results show that while foundation models achieve strong performance in certain settings, simpler models often remain competitive, particularly under clinical distribution shifts. To facilitate reproducibility and adoption, we release all prepared data and code in an accessible and extensible format.

In the process of patient diagnosis, non-invasive measurements are widely used due to their low risks and quick results. Electrocardiogram (ECG), as a non-invasive method to collect heart activities, is used to diagnose cardiac conditions. Analyzing the ECG typically requires domain expertise, which is a roadblock to applying artificial intelligence (AI) for healthcare. Through advances in self-supervised learning and foundation models, AI systems can now acquire and leverage domain knowledge without relying solely on human expertise. However, there is a lack of comprehensive analyses over the foundation models' performance on ECG. This study aims to answer the research question: "Are Foundation Models Useful for ECG Analysis?" To address it, we evaluate language/general time-series/ECG foundation models in comparison with time-series deep learning models. The experimental results show that general time-series/ECG foundation models achieve a top performance rate of 80%, indicating their effectiveness in ECG analysis. In-depth analyses and insights are provided along with comprehensive experimental results. This study highlights the limitations and potential of foundation models in advancing physiological waveform analysis. The data and code for this benchmark are publicly available at https://github.com/yuhaoxu99/ECGMultitasks-Benchmark.

AI based mental health diagnosis is often judged by benchmark accuracy, yet in practice its value depends on how psychologists respond whether they accept, adjust, or reject AI suggestions. Mental health makes this especially challenging: decisions are continuous and shaped by cues in tone, pauses, word choice, and nonverbal behaviors of patients. Current research rarely examines how AI diagnosis interface design influences these choices, leaving little basis for reliable testing before live studies. We present SimClinician, an interactive simulation platform, to transform patient data into psychologist AI collaborative diagnosis. Contributions include: (1) a dashboard integrating audio, text, and gaze-expression patterns; (2) an avatar module rendering de-identified dynamics for analysis; (3) a decision layer that maps AI outputs to multimodal evidence, letting psychologists review AI reasoning, and enter a diagnosis. Tested on the E-DAIC corpus (276 clinical interviews, expanded to 480,000 simulations), SimClinician shows that a confirmation step raises acceptance by 23%, keeping escalations below 9%, and maintaining smooth interaction flow.

Testing humanoid robots with users is slow, causes wear, and limits iteration and diversity. Yet screening agents must master conversational timing, prosody, backchannels, and what to attend to in faces and speech for Depression and PTSD. Most simulators omit policy learning with nonverbal dynamics; many controllers chase task accuracy while underweighting trust, pacing, and rapport. We virtualise the humanoid as a conversational agent to train without hardware burden. Our agent-centred, simulation-first pipeline turns interview data into 276 Unreal Engine MetaHuman patients with synchronised speech, gaze/face, and head-torso poses, plus PHQ-8 and PCL-C flows. A perception-fusion-policy loop decides what and when to speak, when to backchannel, and how to avoid interruptions, under a safety shield. Training uses counterfactual replay (bounded nonverbal perturbations) and an uncertainty-aware turn manager that probes to reduce diagnostic ambiguity. Results are simulation-only; the humanoid is the transfer target. In comparing three controllers, a custom TD3 (Twin Delayed DDPG) outperformed PPO and CEM, achieving near-ceiling coverage with steadier pace at comparable rewards. Decision-quality analyses show negligible turn overlap, aligned cut timing, fewer clarification prompts, and shorter waits. Performance stays stable under modality dropout and a renderer swap, and rankings hold on a held-out patient split. Contributions: (1) an agent-centred simulator that turns interviews into 276 interactive patients with bounded nonverbal counterfactuals; (2) a safe learning loop that treats timing and rapport as first-class control variables; (3) a comparative study (TD3 vs PPO/CEM) with clear gains in completeness and social timing; and (4) ablations and robustness analyses explaining the gains and enabling clinician-supervised humanoid pilots.

Large Language Models (LLMs) may one day replace search engines as the primary portal to information on the Web. In this article, we investigate the societal challenges that such a change could bring. We focus on the roles of LLM Providers, Content Creators, and End Users, and identify 15 types of challenges. With each, we show current mitigation strategies -- both from the technical perspective and the legal perspective. We also discuss the impact of each challenge and point out future research opportunities.

Background: Mentalization integrates cognitive, affective, and intersubjective components. Large Language Models (LLMs) display an increasing ability to generate reflective texts, raising questions regarding the relationship between linguistic form and mental representation. This study assesses the extent to which a single LLM can reproduce the linguistic structure of mentalization according to the parameters of Mentalization-Based Treatment (MBT). Methods: Fifty dialogues were generated between human participants and an LLM configured in standard mode. Five psychiatrists trained in MBT, working under blinded conditions, evaluated the mentalization profiles produced by the model along the four MBT axes, assigning Likert-scale scores for evaluative coherence, argumentative coherence, and global quality. Inter-rater agreement was estimated using ICC(3,1). Results: Mean scores (3.63-3.98) and moderate standard deviations indicate a high level of structural coherence in the generated profiles. ICC values (0.60-0.84) show substantial-to-high agreement among raters. The model proved more stable in the Implicit-Explicit and Self-Other dimensions, while presenting limitations in the integration of internal states and external contexts. The profiles were coherent and clinically interpretable yet characterized by affective neutrality.

Junior indie game developers in distributed, part-time teams lack production frameworks suited to their specific context, as traditional methodologies are often inaccessible. This study introduces the CIGDI (Co-Intelligence Game Development Ideation) Framework, an alternative approach for integrating AI tools to address persistent challenges of technical debt, coordination, and burnout. The framework emerged from a three-month reflective practice and autoethnographic study of a three-person distributed team developing the 2D narrative game "The Worm's Memoirs". Based on analysis of development data (N=157 Jira tasks, N=333 GitHub commits, N=13+ Miro boards, N=8 reflection sessions), CIGDI is proposed as a seven-stage iterative process structured around human-in-the-loop decision points (Priority Criteria and Timeboxing). While AI support democratized knowledge access and reduced cognitive load, our analysis identified a significant challenge: "comprehension debt." We define this as a novel form of technical debt where AI helps teams build systems more sophisticated than their independent skill level can create or maintain. This paradox (possessing functional systems the team incompletely understands) creates fragility and AI dependency, distinct from traditional code quality debt. This work contributes a practical production framework for resource-constrained teams and identifies critical questions about whether AI assistance constitutes a learning ladder or a dependency trap for developer skill.