Archives AI News

Locus: Agentic Predicate Synthesis for Directed Fuzzing

arXiv:2508.21302v2 Announce Type: replace-cross Abstract: Directed fuzzing aims to find program inputs that lead to specified target program states. It has broad applications, such as debugging system crashes, confirming reported bugs, and generating exploits for potential vulnerabilities. This task is inherently challenging because target states are often deeply nested in the program, while the search space manifested by numerous possible program inputs is prohibitively large. Existing approaches rely on branch distances or manually-specified constraints to guide the search; however, the branches alone are often insufficient to precisely characterize progress toward reaching the target states, while the manually specified constraints are often tailored for specific bug types and thus difficult to generalize to diverse target states and programs. We present Locus, a novel framework to improve the efficiency of directed fuzzing. Our key insight is to synthesize predicates to capture fuzzing progress as semantically meaningful intermediate states, serving as milestones towards reaching the target states. When used to instrument the program under fuzzing, they can reject executions unlikely to reach the target states, while providing additional coverage guidance. To automate this task and generalize to diverse programs, Locus features an agentic framework with program analysis tools to synthesize and iteratively refine the candidate predicates, while ensuring the predicates strictly relax the target states to prevent false rejections via symbolic execution. Our evaluation shows that Locus substantially improves the efficiency of eight state-of-the-art fuzzers in discovering real-world vulnerabilities, achieving an average speedup of 41.6x. So far, Locus has found eight previously unpatched bugs, with one already acknowledged with a draft patch.

Synthetic Founders: AI-Generated Social Simulations for Startup Validation Research in Computational Social Science

arXiv:2509.02605v1 Announce Type: cross Abstract: We present a comparative docking experiment that aligns human-subject interview data with large language model (LLM)-driven synthetic personas to evaluate fidelity, divergence, and blind spots in AI-enabled simulation. Fifteen early-stage startup founders were interviewed about their hopes and concerns regarding AI-powered validation, and the same protocol was replicated with AI-generated founder and investor personas. A structured thematic synthesis revealed four categories of outcomes: (1) Convergent themes - commitment-based demand signals, black-box trust barriers, and efficiency gains were consistently emphasized across both datasets; (2) Partial overlaps - founders worried about outliers being averaged away and the stress of real customer validation, while synthetic personas highlighted irrational blind spots and framed AI as a psychological buffer; (3) Human-only themes - relational and advocacy value from early customer engagement and skepticism toward moonshot markets; and (4) Synthetic-only themes - amplified false positives and trauma blind spots, where AI may overstate adoption potential by missing negative historical experiences. We interpret this comparative framework as evidence that LLM-driven personas constitute a form of hybrid social simulation: more linguistically expressive and adaptable than traditional rule-based agents, yet bounded by the absence of lived history and relational consequence. Rather than replacing empirical studies, we argue they function as a complementary simulation category - capable of extending hypothesis space, accelerating exploratory validation, and clarifying the boundaries of cognitive realism in computational social science.

BadPromptFL: A Novel Backdoor Threat to Prompt-based Federated Learning in Multimodal Models

arXiv:2508.08040v2 Announce Type: replace-cross Abstract: Prompt-based tuning has emerged as a lightweight alternative to full fine-tuning in large vision-language models, enabling efficient adaptation via learned contextual prompts. This paradigm has recently been extended to federated learning settings (e.g., PromptFL), where clients collaboratively train prompts under data privacy constraints. However, the security implications of prompt-based aggregation in federated multimodal learning remain largely unexplored, leaving a critical attack surface unaddressed. In this paper, we introduce textbf{BadPromptFL}, the first backdoor attack targeting prompt-based federated learning in multimodal contrastive models. In BadPromptFL, compromised clients jointly optimize local backdoor triggers and prompt embeddings, injecting poisoned prompts into the global aggregation process. These prompts are then propagated to benign clients, enabling universal backdoor activation at inference without modifying model parameters. Leveraging the contextual learning behavior of CLIP-style architectures, BadPromptFL achieves high attack success rates (e.g., (>90%)) with minimal visibility and limited client participation. Extensive experiments across multiple datasets and aggregation protocols validate the effectiveness, stealth, and generalizability of our attack, raising critical concerns about the robustness of prompt-based federated learning in real-world deployments.

MIDOG 2025: Mitotic Figure Detection with Attention-Guided False Positive Correction

arXiv:2509.02598v1 Announce Type: cross Abstract: We present a novel approach which extends the existing Fully Convolutional One-Stage Object Detector (FCOS) for mitotic figure detection. Our composite model adds a Feedback Attention Ladder CNN (FAL-CNN) model for classification of normal versus abnormal mitotic figures, feeding into a fusion network that is trained to generate adjustments to bounding boxes predicted by FCOS. Our network aims to reduce the false positive rate of the FCOS object detector, to improve the accuracy of object detection and enhance the generalisability of the network. Our model achieved an F1 score of 0.655 for mitosis detection on the preliminary evaluation dataset.

Rethinking Data Protection in the (Generative) Artificial Intelligence Era

arXiv:2507.03034v4 Announce Type: replace-cross Abstract: The (generative) artificial intelligence (AI) era has profoundly reshaped the meaning and value of data. No longer confined to static content, data now permeates every stage of the AI lifecycle from the training samples that shape model parameters to the prompts and outputs that drive real-world model deployment. This shift renders traditional notions of data protection insufficient, while the boundaries of what needs safeguarding remain poorly defined. Failing to safeguard data in AI systems can inflict societal and individual, underscoring the urgent need to clearly delineate the scope of and rigorously enforce data protection. In this perspective, we propose a four-level taxonomy, including non-usability, privacy preservation, traceability, and deletability, that captures the diverse protection needs arising in modern (generative) AI models and systems. Our framework offers a structured understanding of the trade-offs between data utility and control, spanning the entire AI pipeline, including training datasets, model weights, system prompts, and AI-generated content. We analyze representative technical approaches at each level and reveal regulatory blind spots that leave critical assets exposed. By offering a structured lens to align future AI technologies and governance with trustworthy data practices, we underscore the urgency of rethinking data protection for modern AI techniques and provide timely guidance for developers, researchers, and regulators alike.

OpenAIs HealthBench in Action: Evaluating an LLM-Based Medical Assistant on Realistic Clinical Queries

arXiv:2509.02594v1 Announce Type: cross Abstract: Evaluating large language models (LLMs) on their ability to generate high-quality, accurate, situationally aware answers to clinical questions requires going beyond conventional benchmarks to assess how these systems behave in complex, high-stake clincal scenarios. Traditional evaluations are often limited to multiple-choice questions that fail to capture essential competencies such as contextual reasoning, awareness and uncertainty handling etc. To address these limitations, we evaluate our agentic, RAG-based clinical support assistant, DR.INFO, using HealthBench, a rubric-driven benchmark composed of open-ended, expert-annotated health conversations. On the Hard subset of 1,000 challenging examples, DR.INFO achieves a HealthBench score of 0.51, substantially outperforming leading frontier LLMs (GPT-5, o3, Grok 3, GPT-4, Gemini 2.5, etc.) across all behavioral axes (accuracy, completeness, instruction following, etc.). In a separate 100-sample evaluation against similar agentic RAG assistants (OpenEvidence, Pathway.md), it maintains a performance lead with a health-bench score of 0.54. These results highlight DR.INFOs strengths in communication, instruction following, and accuracy, while also revealing areas for improvement in context awareness and completeness of a response. Overall, the findings underscore the utility of behavior-level, rubric-based evaluation for building a reliable and trustworthy AI-enabled clinical support assistant.

LLM Embedding-based Attribution (LEA): Quantifying Source Contributions to Generative Model’s Response for Vulnerability Analysis

arXiv:2506.12100v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) are increasingly used for cybersecurity threat analysis, but their deployment in security-sensitive environments raises trust and safety concerns. With over 21,000 vulnerabilities disclosed in 2025, manual analysis is infeasible, making scalable and verifiable AI support critical. When querying LLMs, dealing with emerging vulnerabilities is challenging as they have a training cut-off date. While Retrieval-Augmented Generation (RAG) can inject up-to-date context to alleviate the cut-off date limitation, it remains unclear how much LLMs rely on retrieved evidence versus the model's internal knowledge, and whether the retrieved information is meaningful or even correct. This uncertainty could mislead security analysts, mis-prioritize patches, and increase security risks. Therefore, this work proposes LLM Embedding-based Attribution (LEA) to analyze the generated responses for vulnerability exploitation analysis. More specifically, LEA quantifies the relative contribution of internal knowledge vs. retrieved content in the generated responses. We evaluate LEA on 500 critical vulnerabilities disclosed between 2016 and 2025, across three RAG settings -- valid, generic, and incorrect -- using three state-of-the-art LLMs. Our results demonstrate LEA's ability to detect clear distinctions between non-retrieval, generic-retrieval, and valid-retrieval scenarios with over 95% accuracy on larger models. Finally, we demonstrate the limitations posed by incorrect retrieval of vulnerability information and raise a cautionary note to the cybersecurity community regarding the blind reliance on LLMs and RAG for vulnerability analysis. LEA offers security analysts with a metric to audit RAG-enhanced workflows, improving the transparent and trustworthy deployment of AI in cybersecurity threat analysis.

Robust Pan-Cancer Mitotic Figure Detection with YOLOv12

arXiv:2509.02593v1 Announce Type: cross Abstract: Mitotic figures represent a key histoprognostic feature in tumor pathology, providing crucial insights into tumor aggressiveness and proliferation. However, their identification remains challenging, subject to significant inter-observer variability, even among experienced pathologists. To address this issue, the MItosis DOmain Generalization (MIDOG) 2025 challenge marks the third edition of an international competition aiming to develop robust mitosis detection algorithms. In this paper, we present a mitotic figures detection approach based on the YOLOv12 object detection architecture, achieving a $F_1$-score of 0.801 on the preliminary test set of the MIDOG 2025 challenge, without relying on external data.

Group-in-Group Policy Optimization for LLM Agent Training

arXiv:2505.10978v2 Announce Type: replace-cross Abstract: Recent advances in group-based reinforcement learning (RL) have driven frontier large language models (LLMs) in single-turn tasks like mathematical reasoning. However, their scalability to long-horizon LLM agent training remains limited. Unlike static tasks, agent-environment interactions unfold over many steps and often yield sparse or delayed rewards, making credit assignment across individual steps significantly more challenging. In this work, we propose Group-in-Group Policy Optimization (GiGPO), a novel RL algorithm that achieves fine-grained credit assignment for LLM agents while preserving the appealing properties of group-based RL: critic-free, low memory, and stable convergence. GiGPO introduces a two-level structure for estimating relative advantage: (i) At the episode-level, GiGPO computes macro relative advantages based on groups of complete trajectories; (ii) At the step-level, GiGPO introduces an anchor state grouping mechanism that retroactively constructs step-level groups by identifying repeated environment states across trajectories. Actions stemming from the same state are grouped together, enabling micro relative advantage estimation. This hierarchical structure effectively captures both global trajectory quality and local step effectiveness without relying on auxiliary models or additional rollouts. We evaluate GiGPO on two challenging agent benchmarks, ALFWorld and WebShop, using Qwen2.5-1.5B-Instruct and Qwen2.5-7B-Instruct. Crucially, GiGPO delivers fine-grained per-step credit signals and achieves performance gains of > 12% on ALFWorld and > 9% on WebShop over the GRPO baseline: all while maintaining the same GPU memory overhead, identical LLM rollout, and incurring little to no additional time cost.

Impoola: The Power of Average Pooling for Image-Based Deep Reinforcement Learning

arXiv:2503.05546v2 Announce Type: replace-cross Abstract: As image-based deep reinforcement learning tackles more challenging tasks, increasing model size has become an important factor in improving performance. Recent studies achieved this by focusing on the parameter efficiency of scaled networks, typically using Impala-CNN, a 15-layer ResNet-inspired network, as the image encoder. However, while Impala-CNN evidently outperforms older CNN architectures, potential advancements in network design for deep reinforcement learning-specific image encoders remain largely unexplored. We find that replacing the flattening of output feature maps in Impala-CNN with global average pooling leads to a notable performance improvement. This approach outperforms larger and more complex models in the Procgen Benchmark, particularly in terms of generalization. We call our proposed encoder model Impoola-CNN. A decrease in the network's translation sensitivity may be central to this improvement, as we observe the most significant gains in games without agent-centered observations. Our results demonstrate that network scaling is not just about increasing model size - efficient network design is also an essential factor. We make our code available at https://github.com/raphajaner/impoola.