Detecting vulnerabilities in large-scale, multi-file software systems remains a critical challenge, as traditional techniques and current large language models (LLMs) struggle with long code dependencies and complex control flows. This challenge is further compounded for Java, which dominates enterprise software yet remains relatively underexplored in vulnerability detection research. In this study, we first fine-tune a vulnerability-specific LLM called VulDetLLM and then propose a hybrid vulnerability detection method using VulDetLLM and programassisted language (PAL) model for large-scale source code (HyVD-VP). HyVD-VP performs static detection using semantically meaningful slices and external knowledge, verifies predictions through lightweight dynamic checks during runtime, and integrates static and dynamic signals via an LLM-based decision module to achieve accurate and explainable detection. We evaluate HyVD-VP primarily on Java, which is widely used in missioncritical enterprise domains such as banking, telecommunications, and large-scale web services. The method consistently outperforms traditional analyzers (Fortify, SpotBugs) and recent research baselines, achieving 96.3 % accuracy, 95.7 % F1-score while reducing false negatives. Importantly, it also identified 16 previously unknown vulnerabilities in six real-world enterprise projects, underscoring its industrial relevance. Based on our current evaluation in Java, the detection pipeline shows moderate runtime overhead mainly from dynamic validation, as indicated by runtime profiling. A feasibility study further suggests the potential adaptability of HyVD-VP to multiple programming languages and constrained hardware scenarios simultaneously, indicating both practical applicability and extensibility. These results establish HyVD-VP as a promising step toward scalable, industry-ready solutions for reliable vulnerability detection in large-scale software systems.
When continual test-time adaptation (TTA) persists over the long term, errors accumulate in the model and further cause it to predict only a few classes for all inputs, a phenomenon known as model collapse. Recent studies have explored reset strategies that completely erase these accumulated errors. However, their periodic resets lead to suboptimal adaptation, as they occur independently of the actual risk of collapse. Moreover, their full resets cause catastrophic loss of knowledge acquired over time, even though such knowledge could be beneficial in the future. To this end, we propose (1) an Adaptive and Selective Reset (ASR) scheme that dynamically determines when and where to reset, (2) an importance-aware regularizer to recover essential knowledge lost due to reset, and (3) an on-the-fly adaptation adjustment scheme to enhance adaptability under challenging domain shifts. Extensive experiments across long-term TTA benchmarks demonstrate the effectiveness of our approach, particularly under challenging conditions. Our code is available at https://github.com/YonseiML/asr.
Test-time prompt tuning (TPT) has emerged as a promising technique for enhancing the adaptability of vision-language models by optimizing textual prompts using unlabeled test data. However, prior studies have revealed that TPT often produces poorly calibrated models, raising concerns about the reliability of their predictions. Recent works address this issue by incorporating additional regularization terms that constrain model outputs, which improve calibration but often degrade performance. In this work, we reveal that these regularization strategies implicitly encourage optimization toward flatter minima, and that the sharpness of the loss landscape around adapted prompts is a key factor governing calibration quality. Motivated by this observation, we introduce Flatness-aware Prompt Pretraining (FPP), a simple yet effective pretraining framework for TPT that initializes prompts within flatter regions of the loss landscape prior to adaptation. We show that simply replacing the initialization in existing TPT pipelines---without modifying any other components---is sufficient to improve both calibration and performance. Notably, FPP requires no labeled data and avoids any additional computational costs during test-time tuning, making it highly practical for real-world deployment. The code will be released.
Conventional robot social behavior generation has been limited in flexibility and autonomy, relying on predefined motions or human feedback. This study proposes CRISP (Critique-and-Replan for Interactive Social Presence), an autonomous framework where a robot critiques and replans its own actions by leveraging a Vision-Language Model (VLM) as a `human-like social critic.' CRISP integrates (1) extraction of movable joints and constraints by analyzing the robot's description file (e.g., MJCF), (2) generation of step-by-step behavior plans based on situational context, (3) generation of low-level joint control code by referencing visual information (joint range-of-motion visualizations), (4) VLM-based evaluation of social appropriateness and naturalness, including pinpointing erroneous steps, and (5) iterative refinement of behaviors through reward-based search. This approach is not tied to a specific robot API; it can generate subtly different, human-like motions on various platforms using only the robot's structure file. In a user study involving five different robot types and 20 scenarios, including mobile manipulators and humanoids, our proposed method achieved significantly higher preference and situational appropriateness ratings compared to previous methods. This research presents a general framework that minimizes human intervention while expanding the robot's autonomous interaction capabilities and cross-platform applicability. Detailed result videos and supplementary information regarding this work are available at: https://limjiyu99.github.io/inner-critic/
Federated learning (FL) over wireless networks has become a key enabler for privacy-preserving distributed artificial intelligence (AI). However, high learning latency remains a critical bottleneck due to the presence of stragglers, limited wireless resources, and frequent model uploads. While model quantization can mitigate this issue by reducing communication overhead, its effectiveness is sensitive to device heterogeneity and time-varying channel conditions. To address this issue, we propose FedDamQu, a communication-efficient FL framework with doubly adaptive model quantization, which dynamically adjusts quantization bit widths across devices and communication rounds to balance latency and accuracy. Our objective is to maximize the model performance under learning latency constraints. The main contributions are summarized as follows: 1) convergence analysis under unreliable channels: we derive a novel convergence error upper bound for FedDamQu, which explicitly quantifies the impact of device selection, unreliable transmission, and quantization error on the global model performance, under both fixed and dynamic quantization gain (QG) settings; 2) joint optimization framework: Based on the knowledge from the proposed theoretical bound, we formulate a joint mixed-integer nonlinear programming (MINLP) problem that integrates device selection, quantization bit-width configuration, and bandwidth allocation to minimize the convergence error under latency constraints; and 3) efficient solution design: the MINLP problem is decomposed into three subproblems, where closed-form solutions for quantization bit-width configuration and bandwidth allocation subproblems are derived, and a lightweight yet effective iterative algorithm is developed to obtain a suboptimal solution for the device selection subproblem. The extensive experiment results validate the theoretical analysis and demonstrate that FedDamQu consistently outperforms existing methods in terms of convergence rate and model accuracy, while significantly reducing the overall learning latency.
