Large multimodal models (LMMs) have advanced significantly by integrating visual encoders with extensive language models, enabling robust reasoning capabilities. However, compressing LMMs for deployment on edge devices remains a critical challenge. In this work, we propose an adaptive search algorithm that optimizes sparsity and KV cache compression to enhance LMM efficiency. Utilizing the Tree-structured Parzen Estimator, our method dynamically adjusts pruning ratios and KV cache quantization bandwidth across different LMM layers, using model performance as the optimization objective. This approach uniquely combines pruning with key-value cache quantization and incorporates a fast pruning technique that eliminates the need for additional fine-tuning or weight adjustments, achieving efficient compression without compromising accuracy. Comprehensive evaluations on benchmark datasets, including LLaVA-1.5 7B and 13B, demonstrate our method superiority over state-of-the-art techniques such as SparseGPT and Wanda across various compression levels. Notably, our framework automatic allocation of KV cache compression resources sets a new standard in LMM optimization, delivering memory efficiency without sacrificing much performance.

Neural Architecture Search (NAS) is challenged by the tradeoff between search space exploration and efficiency, especially for complex tasks. While recent LLM-based NAS methods have shown promise, they often suffer from static search strategies and ambiguous architecture representations. We propose PhaseNAS, an LLM-based NAS framework with dynamic phase transitions guided by real-time score thresholds and a structured architecture template language for consistent code generation. On the NAS-Bench-Macro benchmark, PhaseNAS consistently discovers architectures with higher accuracy and better rank. For image classification (CIF AR-10/100), PhaseNAS reduces search time by up to 86% while maintaining or improving accuracy. In object detection, it automatically produces YOLOv8 variants with higher mAP and lower resource cost. These results demonstrate that PhaseNAS enables efficient, adaptive, and generalizable NAS across diverse vision tasks.

With the rapid advancement of generative models, associated privacy concerns have attracted growing attention. To address this, researchers have begun adapting machine unlearning techniques from traditional classification models to generative settings. Although notable progress has been made in this area, a unified framework for systematically organizing and integrating existing work is still lacking. The substantial differences among current studies in terms of unlearning objectives and evaluation protocols hinder the objective and fair comparison of various approaches. While some studies focus on specific types of generative models, they often overlook the commonalities and systematic characteristics inherent in Generative Model Unlearning (GenMU). To bridge this gap, we provide a comprehensive review of current research on GenMU and propose a unified analytical framework for categorizing unlearning objectives, methodological strategies, and evaluation metrics. In addition, we explore the connections between GenMU and related techniques, including model editing, reinforcement learning from human feedback, and controllable generation. We further highlight the potential practical value of unlearning techniques in real-world applications. Finally, we identify key challenges and outline future research directions aimed at laying a solid foundation for further advancements in this field. We consistently maintain the related open-source materials at https://github.com/caxLee/Generative-model-unlearning-survey.

MSC 804 5 - 0020 - 10 St. Louis, MO 63110 Financial Support: P30 - AR073752 Conflict of interest: No conflicts declared. Page 2 of 13 Roberson 2025 - LLMs in medicine Abstract The medical ecosystem consists of the training of new clinicians and researchers, the practice of clinical medicine, and areas of adjacen t research. There are many aspects of these domains that could benefit from the application of task automation and programmatic assistance. Machine learning and artificial intelligence techniques, including large language models ( LLMs), have been promised to deliver on healthcare innovation, improving care speed and accuracy, and reducing the burden on staff for manual interventions . However, LLMs have no understanding of objective truth that is based in reality. They also represent real risks to the disclo sure of protected information when used by clinicians and researchers. The use of AI in medicine in general, and the deployment of LLMs in particular, therefore requires careful consideration and thoughtful application to reap the benefits of these technol ogies while avoiding the dangers in each context. Page 3 of 13 Roberson 2025 - LLMs in medicine Abbreviations AI: Artificial intelligence CSP: Constrained solution problems EHR: Electronic health records GPT: Generative pre - trained transformer LLM: Large language model ML: Machine learning RAG: Retrieval - augmented generation Page 4 of 13 Roberson 2025 - LLMs in medicine What is a large language model? Large language models are one of the most hyped artificial intelligence technologies of the past few years. Buzz terms associated with them include artificial intelligence, machine learning, a nd deep learning.

Root Cause Analysis (RCA) in telecommunication networks is a critical task, yet it presents a formidable challenge for Artificial Intelligence (AI) due to its complex, graph-based reasoning requirements and the scarcity of realistic benchmarks. To catalyze research in this domain, We herein present TN-RCA530, the inaugural real-world, publicly accessible benchmark for root cause analysis (RCA) of telecommunication network alarms, comprising 530 fault scenarios constructed from expert-validated Knowledge Graphs(KGs). Our evaluation reveals that even state-of-the-art Large Language Models (LLMs) perform poorly on this task, with the best models achieving an F1-score below 70%, highlighting its significant difficulty.To address this challenge, we then propose Auto-RCA, a novel agentic system that automates the iterative refinement of a code-based solution. The core innovation of Auto-RCA lies beyond simple self-correction; it employs an iterative "evaluate-analyze-repair" loop that systematically identifies common patterns across all failure cases to generate contrastive feedback. This feedback guides the LLM to fix systemic logical flaws rather than isolated errors. Experiments show that this agentic framework dramatically boosts problem-solving performance, elevating the final solution's F1-score on TN-RCA530 from a baseline of 58.99% (achieved by Gemini-2.5-Pro

This paper introduces Group Sequence Policy Optimization (GSPO), our stable, efficient, and performant reinforcement learning algorithm for training large language models. Unlike previous algorithms that adopt token-level importance ratios, GSPO defines the importance ratio based on sequence likelihood and performs sequence-level clipping, rewarding, and optimization. We demonstrate that GSPO achieves superior training efficiency and performance compared to the GRPO algorithm, notably stabilizes Mixture-of-Experts (MoE) RL training, and has the potential for simplifying the design of RL infrastructure. These merits of GSPO have contributed to the remarkable improvements in the latest Qwen3 models.

Graphical User Interface (GUI) grounding maps natural language instructions to precise interface locations for autonomous interaction. Current reinforcement learning approaches use binary rewards that treat elements as hit-or-miss targets, creating sparse signals that ignore the continuous nature of spatial interactions. Motivated by human clicking behavior that naturally forms Gaussian distributions centered on target elements, we introduce GUI Gaussian Grounding Rewards (GUI-G$^2$), a principled reward framework that models GUI elements as continuous Gaussian distributions across the interface plane. GUI-G$^2$ incorporates two synergistic mechanisms: Gaussian point rewards model precise localization through exponentially decaying distributions centered on element centroids, while coverage rewards assess spatial alignment by measuring the overlap between predicted Gaussian distributions and target regions. To handle diverse element scales, we develop an adaptive variance mechanism that calibrates reward distributions based on element dimensions. This framework transforms GUI grounding from sparse binary classification to dense continuous optimization, where Gaussian distributions generate rich gradient signals that guide models toward optimal interaction positions. Extensive experiments across ScreenSpot, ScreenSpot-v2, and ScreenSpot-Pro benchmarks demonstrate that GUI-G$^2$, substantially outperforms state-of-the-art method UI-TARS-72B, with the most significant improvement of 24.7% on ScreenSpot-Pro. Our analysis reveals that continuous modeling provides superior robustness to interface variations and enhanced generalization to unseen layouts, establishing a new paradigm for spatial reasoning in GUI interaction tasks.

The use of large language models (LLMs) in hiring promises to streamline candidate screening, but it also raises serious concerns regarding accuracy and algorithmic bias where sufficient safeguards are not in place. In this work, we benchmark several state-of-the-art foundational LLMs - including models from OpenAI, Anthropic, Google, Meta, and Deepseek, and compare them with our proprietary domain-specific hiring model (Match Score) for job candidate matching. We evaluate each model's predictive accuracy (ROC AUC, Precision-Recall AUC, F1-score) and fairness (impact ratio of cut-off analysis across declared gender, race, and intersectional subgroups). Our experiments on a dataset of roughly 10,000 real-world recent candidate-job pairs show that Match Score outperforms the general-purpose LLMs on accuracy (ROC AUC 0.85 vs 0.77) and achieves significantly more equitable outcomes across demographic groups. Notably, Match Score attains a minimum race-wise impact ratio of 0.957 (near-parity), versus 0.809 or lower for the best LLMs, (0.906 vs 0.773 for the intersectionals, respectively). We discuss why pretraining biases may cause LLMs with insufficient safeguards to propagate societal biases in hiring scenarios, whereas a bespoke supervised model can more effectively mitigate these biases. Our findings highlight the importance of domain-specific modeling and bias auditing when deploying AI in high-stakes domains such as hiring, and caution against relying on off-the-shelf LLMs for such tasks without extensive fairness safeguards. Furthermore, we show with empirical evidence that there shouldn't be a dichotomy between choosing accuracy and fairness in hiring: a well-designed algorithm can achieve both accuracy in hiring and fairness in outcomes.

Existing threat modeling frameworks related to transportation cyber-physical systems (CPS) are often narrow in scope, labor-intensive, and require substantial cybersecurity expertise. To this end, we introduce the Transportation Cybersecurity and Resiliency Threat Modeling Framework (TraCR-TMF), a large language model (LLM)-based threat modeling framework for transportation CPS that requires limited cybersecurity expert intervention. TraCR-TMF identifies threats, potential attack techniques, and relevant countermeasures for transportation CPS. Three LLM-based approaches support these identifications: (i) a retrieval-augmented generation approach requiring no cybersecurity expert intervention, (ii) an in-context learning approach with low expert intervention, and (iii) a supervised fine-tuning approach with moderate expert intervention. TraCR-TMF offers LLM-based attack path identification for critical assets based on vulnerabilities across transportation CPS entities. Additionally, it incorporates the Common Vulnerability Scoring System (CVSS) scores of known exploited vulnerabilities to prioritize threat mitigations. The framework was evaluated through two cases. First, the framework identified relevant attack techniques for various transportation CPS applications, 73% of which were validated by cybersecurity experts as correct. Second, the framework was used to identify attack paths for a target asset in a real-world cyberattack incident. TraCR-TMF successfully predicted exploitations, like lateral movement of adversaries, data exfiltration, and data encryption for ransomware, as reported in the incident. These findings show the efficacy of TraCR-TMF in transportation CPS threat modeling, while reducing the need for extensive involvement of cybersecurity experts. To facilitate real-world adoptions, all our codes are shared via an open-source repository.

Psychology research has shown that humans are poor at estimating their performance on tasks, tending towards underconfidence on easy tasks and overconfidence on difficult tasks. We examine three LLMs, Llama-3-70B-instruct, Claude-3-Sonnet, and GPT-4o, on a range of QA tasks of varying difficulty, and show that models exhibit subtle differences from human patterns of overconfidence: less sensitive to task difficulty, and when prompted to answer based on different personas -- e.g., expert vs layman, or different race, gender, and ages -- the models will respond with stereotypically biased confidence estimations even though their underlying answer accuracy remains the same. Based on these observations, we propose Answer-Free Confidence Estimation (AFCE) to improve confidence calibration and LLM interpretability in these settings. AFCE is a self-assessment method that employs two stages of prompting, first eliciting only confidence scores on questions, then asking separately for the answer. Experiments on the MMLU and GPQA datasets spanning subjects and difficulty show that this separation of tasks significantly reduces overconfidence and delivers more human-like sensitivity to task difficulty.