Models like OpenAI-o3 pioneer visual grounded reasoning by dynamically referencing visual regions, just like human "thinking with images". However, no benchmark exists to evaluate these capabilities holistically. To bridge this gap, we propose TreeBench (Traceable Evidence Evaluation Benchmark), a diagnostic benchmark built on three principles: (1) focused visual perception of subtle targets in complex scenes, (2) traceable evidence via bounding box evaluation, and (3) second-order reasoning to test object interactions and spatial hierarchies beyond simple object localization. Prioritizing images with dense objects, we initially sample 1K high-quality images from SA-1B, and incorporate eight LMM experts to manually annotate questions, candidate options, and answers for each image. After three stages of quality control, TreeBench consists of 405 challenging visual question-answering pairs, even the most advanced models struggle with this benchmark, where none of them reach 60% accuracy, e.g., OpenAI-o3 scores only 54.87. Furthermore, we introduce TreeVGR (Traceable Evidence Enhanced Visual Grounded Reasoning), a training paradigm to supervise localization and reasoning jointly with reinforcement learning, enabling accurate localizations and explainable reasoning pathways. Initialized from Qwen2.5-VL-7B, it improves V* Bench (+16.8), MME-RealWorld (+12.6), and TreeBench (+13.4), proving traceability is key to advancing vision-grounded reasoning. The code is available at https://github.com/Haochen-Wang409/TreeVGR.

As large language models (LLMs) become increasingly integrated into clinical decision-making, ensuring transparent and trustworthy reasoning is essential. However, existing evaluation strategies of LLMs' medical reasoning capability either suffer from unsatisfactory assessment or poor scalability, and a rigorous benchmark remains lacking. To address this, we introduce MedThink-Bench, a benchmark designed for rigorous, explainable, and scalable assessment of LLMs' medical reasoning. MedThink -Bench comprises 500 challenging questions across ten medical domains, each annotated with expert-crafted step-by -step rationales. Building on this, we propose LLM -w-Ref, a novel evaluation framework that leverages fine-grained rationales and LLM -as -a -Judge mechanisms to assess intermediate reasoning with expert -level fidelity while maintaining scalability. Experiments show that LLM -w-Ref exhibits a strong positive correlation with expert judgments. Benchmarking twelve state-of -the -art LLMs, we find that smaller models (e.g ., MedGemma-27B) can surpass larger proprietary counterparts (e.g., OpenAI -o3). Overall, MedThink-Bench offers a foundational tool for evaluating LLMs' medical reasoning, advancing their safe and responsible deployment in clinical practice.

Transposed Convolutions (TCONV) enable the up-scaling mechanism within generative Artificial Intelligence (AI) models. However, the predominant Input-Oriented Mapping (IOM) method for implementing TCONV has complex output mapping, overlapping sums, and ineffectual computations. These inefficiencies further exacerbate the performance bottleneck of TCONV and generative models on resource-constrained edge devices. To address this problem, in this paper we propose MM2IM, a hardware-software co-designed accelerator that combines Matrix Multiplication (MatMul) with col2IM to process TCONV layers on resource-constrained edge devices efficiently. Using the SECDA-TFLite design toolkit, we implement MM2IM and evaluate its performance across 261 TCONV problem configurations, achieving an average speedup of 1.9x against a dual-thread ARM Neon optimized CPU baseline. We then evaluate the performance of MM2IM on a range of TCONV layers from well-known generative models achieving up to 4.2x speedup, and compare it against similar resource-constrained TCONV accelerators, outperforming them by at least 2x GOPs/DSP. Finally, we evaluate MM2IM on the DCGAN and pix2pix GAN models, achieving up to 3x speedup and 2.4x energy reduction against the CPU baseline.

Large vision-language models (VLMs) increasingly adopt post-training techniques such as long chain-of-thought (CoT) supervised fine-tuning (SFT) and reinforcement learning (RL) to elicit sophisticated reasoning. While these methods exhibit synergy in language-only models, their joint effectiveness in VLMs remains uncertain. We present a systematic investigation into the distinct roles and interplay of long-CoT SFT and RL across multiple multimodal reasoning benchmarks. We find that SFT improves performance on difficult questions by in-depth, structured reasoning, but introduces verbosity and degrades performance on simpler ones. In contrast, RL promotes generalization and brevity, yielding consistent improvements across all difficulty levels, though the improvements on the hardest questions are less prominent compared to SFT. Surprisingly, combining them through two-staged, interleaved, or progressive training strategies, as well as data mixing and model merging, all fails to produce additive benefits, instead leading to trade-offs in accuracy, reasoning style, and response length. This "synergy dilemma" highlights the need for more seamless and adaptive approaches to unlock the full potential of combined post-training techniques for reasoning VLMs.Figure 1: Accuracy gains from various post-training techniques across five difficulty levels (L1, easy to L5, hard) on five multimodal reasoning benchmarks. Long-CoT SFT boosts Qwen2.5-VL-7B on harder questions but hurts easier ones, while RL yields steady gains across the board. Hybrid strategies consistently trade off strengths rather than achieving true synergy. Large language models (LLMs) like OpenAI's o1/o3 (Jaech et al., 2024) and DeepSeek-R1 (Guo et al., 2025) have demonstrated remarkable reasoning abilities by thinking before answering .

What algorithms do LLMs actually learn and use to solve problems? Studies addressing this question are sparse, as research priorities are focused on improving performance through scale, leaving a theoretical and empirical gap in understanding emergent algorithms. This position paper proposes AlgEval: a framework for systematic research into the algorithms that LLMs learn and use. AlgEval aims to uncover algorithmic primitives, reflected in latent representations, attention, and inference-time compute, and their algorithmic composition to solve task-specific problems. We highlight potential methodological paths and a case study toward this goal, focusing on emergent search algorithms. Our case study illustrates both the formation of top-down hypotheses about candidate algorithms, and bottom-up tests of these hypotheses via circuit-level analysis of attention patterns and hidden states. The rigorous, systematic evaluation of how LLMs actually solve tasks provides an alternative to resource-intensive scaling, reorienting the field toward a principled understanding of underlying computations. Such algorithmic explanations offer a pathway to human-understandable interpretability, enabling comprehension of the model's internal reasoning performance measures. This can in turn lead to more sample-efficient methods for training and improving performance, as well as novel architectures for end-to-end and multi-agent systems.

Generative AI has demonstrated strong potential in healthcare, from clinical decision support to patient-facing chatbots that improve outcomes. A critical challenge for deployment is effective human-AI communication, where content must be both personalized and understandable. We introduce MedReadCtrl, a readability-controlled instruction tuning framework that enables LLMs to adjust output complexity without compromising meaning. Evaluations of nine datasets and three tasks across medical and general domains show that MedReadCtrl achieves significantly lower readability instruction-following errors than GPT-4 (e.g., 1.39 vs. 1.59 on ReadMe, p<0.001) and delivers substantial gains on unseen clinical tasks (e.g., +14.7 ROUGE-L, +6.18 SARI on MTSamples). Experts consistently preferred MedReadCtrl (71.7% vs. 23.3%), especially at low literacy levels. These gains reflect MedReadCtrl's ability to restructure clinical content into accessible, readability-aligned language while preserving medical intent, offering a scalable solution to support patient education and expand equitable access to AI-enabled care.

A poster on Reddit lived with a painful clicking jaw, the result of a boxing injury, for five years. They saw specialists, got MRIs, but no one could give them a solution to fix it, until they described the problem to ChatGPT. The AI chatbot suggested a specific jaw-alignment issue might be the problem and offered a technique involving tongue placement as a treatment. The individual tried it, and the clicking stopped. "After five years of just living with it," they wrote on Reddit in April, "this AI gave me a fix in a minute."

Elon Musk on Thursday unveiled Grok 4, the latest AI model from xAI, his multibillion-dollar initiative to rival OpenAI and Google. Without citing detailed evidence, Musk claimed that the model aces standardized tests and exhibits doctorate-level knowledge in a wide array of different disciplines. "Grok 4 is a postgrad-level in everything," Musk said during an hour-long live broadcast, which began after midnight in New York. "At least with respect to academic questions, Grok 4 is better than PhD level in every subject. Competing AI developers, such as OpenAI and Google, have routinely released similar publications for their models.

We introduce our first reflective generative model MetaStone-S1, which obtains OpenAI o3-mini's performance via the new Reflective Generative Form. The new form focuses on high-quality reasoning trajectory selection and contains two novelties: 1) A unified interface for policy and process reward model: we share the backbone network and use task-specific heads for reasoning trajectory predicting and scoring respectively, introducing only 53M extra parameters for trajectory scoring. 2) Eliminating the reliance on process-level annotation: we provide a self-supervised process reward model, which can directly learn the high-quality reasoning trajectory selection from the outcome reward. Equipped with the reflective generative form, MetaStone-S1 is naturally suitable for test-time scaling, and we provide three reasoning effort modes (low, medium, and high) based on the controllable thinking length. Experiments demonstrate that our MetaStone-S1 achieves comparable performance to OpenAI o3-mini's series with only 32B parameter size. To support the research community, we have open-sourced MetaStone-S1 at https://github.com/MetaStone-AI/MetaStone-S1.

Federated learning (FL) has emerged as a promising paradigm for collaboratively training deep learning models across institutions without exchanging sensitive medical data. However, its effectiveness is often hindered by limited data availability and non-independent, identically distributed data across participating clients, which can degrade model performance and generalization. To address these challenges, we propose a generative AI based data augmentation framework that integrates synthetic image sharing into the federated training process for breast cancer diagnosis via ultrasound images. Specifically, we train two simple class-specific Deep Convolutional Generative Adversarial Networks: one for benign and one for malignant lesions. We then simulate a realistic FL setting using three publicly available breast ultrasound image datasets: BUSI, BUS-BRA, and UDIAT. FedAvg and FedProx are adopted as baseline FL algorithms. Experimental results show that incorporating a suitable number of synthetic images improved the average AUC from 0.9206 to 0.9237 for FedAvg and from 0.9429 to 0.9538 for FedProx. We also note that excessive use of synthetic data reduced performance, underscoring the importance of maintaining a balanced ratio of real and synthetic samples. Our findings highlight the potential of generative AI based data augmentation to enhance FL results in the breast ultrasound image classification task.