Yuan Xia

Yuan Xia, Jingbo Zhou, Zhenhui Shi et al.

The Retrieval-Augmented Language Model (RALM) has shown remarkable performance on knowledge-intensive tasks by incorporating external knowledge during inference, which mitigates the factual hallucinations inherited in large language models (LLMs). Despite these advancements, challenges persist in the implementation of RALMs, particularly concerning their reliability and traceability. To be specific, the irrelevant document retrieval may result in unhelpful response generation or even deteriorate the performance of LLMs, while the lack of proper citations in generated outputs complicates efforts to verify the trustworthiness of the models. To this end, we propose a novel self-reasoning framework aimed at improving the reliability and traceability of RALMs, whose core idea is to leverage reasoning trajectories generated by the LLM itself. The framework involves constructing self-reason trajectories with three processes: a relevance-aware process, an evidence-aware selective process, and a trajectory analysis process. We have evaluated our framework across four public datasets (two short-form QA datasets, one long-form QA dataset, and one fact verification dataset) to demonstrate the superiority of our method, which can outperform existing state-of-the-art models and can achieve comparable performance with GPT-4, while only using 2,000 training samples.

Pengfei Luo, Jingbo Zhou, Tong Xu et al.

With the proliferation of images in online content, language-guided image retrieval (LGIR) has emerged as a research hotspot over the past decade, encompassing a variety of subtasks with diverse input forms. While the development of large multimodal models (LMMs) has significantly facilitated these tasks, existing approaches often address them in isolation, requiring the construction of separate systems for each task. This not only increases system complexity and maintenance costs, but also exacerbates challenges stemming from language ambiguity and complex image content, making it difficult for retrieval systems to provide accurate and reliable results. To this end, we propose ImageScope, a training-free, three-stage framework that leverages collective reasoning to unify LGIR tasks. The key insight behind the unification lies in the compositional nature of language, which transforms diverse LGIR tasks into a generalized text-to-image retrieval process, along with the reasoning of LMMs serving as a universal verification to refine the results. To be specific, in the first stage, we improve the robustness of the framework by synthesizing search intents across varying levels of semantic granularity using chain-of-thought (CoT) reasoning. In the second and third stages, we then reflect on retrieval results by verifying predicate propositions locally, and performing pairwise evaluations globally. Experiments conducted on six LGIR datasets demonstrate that ImageScope outperforms competitive baselines. Comprehensive evaluations and ablation studies further confirm the effectiveness of our design.

Yuan Xia, Akanksha Atrey, Fadoua Khmaissia et al.

This paper investigates the logical reasoning capabilities of large language models (LLMs). For a precisely defined yet tractable formulation, we choose the conceptually simple but technically complex task of constructing proofs in Boolean logic. A trained LLM receives as input a set of assumptions and a goal, and produces as output a proof that formally derives the goal from the assumptions. Incorrect proofs are caught by an automated proof checker. A critical obstacle for training is the scarcity of real-world proofs. We propose an efficient, randomized procedure for synthesizing valid proofs and introduce Template Transformation, a data augmentation technique that enhances the model's ability to handle complex logical expressions. The central evaluation question is whether an LLM has indeed learned to reason. We propose tests to measure the reasoning ability of a black-box LLM. By these measures, experiments demonstrate strong reasoning capabilities for assertions with short proofs, which decline with proof complexity. Notably, template transformation improves accuracy even for smaller models, suggesting its effectiveness across model scales.

Amin Adibi, Xu Cao, Zongliang Ji et al.

The fourth Machine Learning for Health (ML4H) symposium was held in person on December 15th and 16th, 2024, in the traditional, ancestral, and unceded territories of the Musqueam, Squamish, and Tsleil-Waututh Nations in Vancouver, British Columbia, Canada. The symposium included research roundtable sessions to foster discussions between participants and senior researchers on timely and relevant topics for the ML4H community. The organization of the research roundtables at the conference involved 13 senior and 27 junior chairs across 13 tables. Each roundtable session included an invited senior chair (with substantial experience in the field), junior chairs (responsible for facilitating the discussion), and attendees from diverse backgrounds with an interest in the session's topic.

Xiuyuan Chen, Tao Sun, Dexin Su et al.

Current benchmarks for AI clinician systems, often based on multiple-choice exams or manual rubrics, fail to capture the depth, robustness, and safety required for real-world clinical practice. To address this, we introduce the GAPS framework, a multidimensional paradigm for evaluating \textbf{G}rounding (cognitive depth), \textbf{A}dequacy (answer completeness), \textbf{P}erturbation (robustness), and \textbf{S}afety. Critically, we developed a fully automated, guideline-anchored pipeline to construct a GAPS-aligned benchmark end-to-end, overcoming the scalability and subjectivity limitations of prior work. Our pipeline assembles an evidence neighborhood, creates dual graph and tree representations, and automatically generates questions across G-levels. Rubrics are synthesized by a DeepResearch agent that mimics GRADE-consistent, PICO-driven evidence review in a ReAct loop. Scoring is performed by an ensemble of large language model (LLM) judges. Validation confirmed our automated questions are high-quality and align with clinician judgment. Evaluating state-of-the-art models on the benchmark revealed key failure modes: performance degrades sharply with increased reasoning depth (G-axis), models struggle with answer completeness (A-axis), and they are highly vulnerable to adversarial perturbations (P-axis) as well as certain safety issues (S-axis). This automated, clinically-grounded approach provides a reproducible and scalable method for rigorously evaluating AI clinician systems and guiding their development toward safer, more reliable clinical practice.

Fangxin Shang, Yuan Xia, Dalu Yang et al.

Medical report interpretation plays a crucial role in healthcare, enabling both patient-facing explanations and effective information flow across clinical systems. While recent vision-language models (VLMs) and large language models (LLMs) have demonstrated general document understanding capabilities, there remains a lack of standardized benchmarks to assess structured interpretation quality in medical reports. We introduce MedRepBench, a comprehensive benchmark built from 1,900 de-identified real-world Chinese medical reports spanning diverse departments, patient demographics, and acquisition formats. The benchmark is designed primarily to evaluate end-to-end VLMs for structured medical report understanding. To enable controlled comparisons, we also include a text-only evaluation setting using high-quality OCR outputs combined with LLMs, allowing us to estimate the upper-bound performance when character recognition errors are minimized. Our evaluation framework supports two complementary protocols: (1) an objective evaluation measuring field-level recall of structured clinical items, and (2) an automated subjective evaluation using a powerful LLM as a scoring agent to assess factuality, interpretability, and reasoning quality. Based on the objective metric, we further design a reward function and apply Group Relative Policy Optimization (GRPO) to improve a mid-scale VLM, achieving up to 6% recall gain. We also observe that the OCR+LLM pipeline, despite strong performance, suffers from layout-blindness and latency issues, motivating further progress toward robust, fully vision-based report understanding.

Zhang Xuerui, Yuan Xia

Objective Semi-supervised video object segmentation refers to segmenting the object in subsequent frames given the object label in the first frame. Existing algorithms are mostly based on the objectives of matching and propagation strategies, which often make use of the previous frame with masking or optical flow. This paper explores a new propagation method, uses short-term matching modules to extract the information of the previous frame and apply it in propagation, and proposes the network of Long-Short-Term similarity matching for video object segmentation (LSMOVS) Method: By conducting pixel-level matching and correlation between long-term matching module and short-term matching module with the first frame and previous frame, global similarity map and local similarity map are obtained, as well as feature pattern of current frame and masking of previous frame. After two refine networks, final results are obtained through segmentation network. Results: According to the experiments on the two data sets DAVIS 2016 and 2017, the method of this paper achieves favorable average of region similarity and contour accuracy without online fine tuning, which achieves 86.5% and 77.4% in terms of single target and multiple targets. Besides, the count of segmented frames per second reached 21. Conclusion: The short-term matching module proposed in this paper is more conducive to extracting the information of the previous frame than only the mask. By combining the long-term matching module with the short-term matching module, the whole network can achieve efficient video object segmentation without online fine tuning