Rui Feng

Xinyu Huang, Youcai Zhang, Jinyu Ma et al. · microsoft-research

This paper presents Tag2Text, a vision language pre-training (VLP) framework, which introduces image tagging into vision-language models to guide the learning of visual-linguistic features. In contrast to prior works which utilize object tags either manually labeled or automatically detected with an off-the-shelf detector with limited performance, our approach explicitly learns an image tagger using tags parsed from image-paired text and thus provides a strong semantic guidance to vision-language models. In this way, Tag2Text can utilize large-scale annotation-free image tags in accordance with image-text pairs, and provides more diverse tag categories beyond objects. As a result, Tag2Text demonstrates the ability of a foundational image tagging model, with superior zero-shot performance even comparable to fully supervised models. Moreover, by leveraging the tagging guidance, Tag2Text effectively enhances the performance of vision-language models on both generation-based and alignment-based tasks. Across a wide range of downstream benchmarks, Tag2Text achieves state-of-the-art results with similar model sizes and data scales, demonstrating the efficacy of the proposed tagging guidance. Code, demo and pre-trained models are available at https://github.com/xinyu1205/recognize-anything.

Jilan Xu, Junlin Hou, Yuejie Zhang et al.

Weakly Supervised Object Localization (WSOL) aims to localize objects with image-level supervision. Existing works mainly rely on Class Activation Mapping (CAM) derived from a classification model. However, CAM-based methods usually focus on the most discriminative parts of an object (i.e., incomplete localization problem). In this paper, we empirically prove that this problem is associated with the mixup of the activation values between less discriminative foreground regions and the background. To address it, we propose Class RE-Activation Mapping (CREAM), a novel clustering-based approach to boost the activation values of the integral object regions. To this end, we introduce class-specific foreground and background context embeddings as cluster centroids. A CAM-guided momentum preservation strategy is developed to learn the context embeddings during training. At the inference stage, the re-activation mapping is formulated as a parameter estimation problem under Gaussian Mixture Model, which can be solved by deriving an unsupervised Expectation-Maximization based soft-clustering algorithm. By simply integrating CREAM into various WSOL approaches, our method significantly improves their performance. CREAM achieves the state-of-the-art performance on CUB, ILSVRC and OpenImages benchmark datasets. Code will be available at https://github.com/Jazzcharles/CREAM.

Junlin Hou, Jilan Xu, Fan Xiao et al.

Automatic diabetic retinopathy (DR) grading based on fundus photography has been widely explored to benefit the routine screening and early treatment. Existing researches generally focus on single-field fundus images, which have limited field of view for precise eye examinations. In clinical applications, ophthalmologists adopt two-field fundus photography as the dominating tool, where the information from each field (i.e.,macula-centric and optic disc-centric) is highly correlated and complementary, and benefits comprehensive decisions. However, automatic DR grading based on two-field fundus photography remains a challenging task due to the lack of publicly available datasets and effective fusion strategies. In this work, we first construct a new benchmark dataset (DRTiD) for DR grading, consisting of 3,100 two-field fundus images. To the best of our knowledge, it is the largest public DR dataset with diverse and high-quality two-field images. Then, we propose a novel DR grading approach, namely Cross-Field Transformer (CrossFiT), to capture the correspondence between two fields as well as the long-range spatial correlations within each field. Considering the inherent two-field geometric constraints, we particularly define aligned position embeddings to preserve relative consistent position in fundus. Besides, we perform masked cross-field attention during interaction to flter the noisy relations between fields. Extensive experiments on our DRTiD dataset and a public DeepDRiD dataset demonstrate the effectiveness of our CrossFiT network. The new dataset and the source code of CrossFiT will be publicly available at https://github.com/FDU-VTS/DRTiD.

Jiashuo Yu, Jinyu Liu, Ying Cheng et al.

Weakly-supervised audio-visual violence detection aims to distinguish snippets containing multimodal violence events with video-level labels. Many prior works perform audio-visual integration and interaction in an early or intermediate manner, yet overlooking the modality heterogeneousness over the weakly-supervised setting. In this paper, we analyze the modality asynchrony and undifferentiated instances phenomena of the multiple instance learning (MIL) procedure, and further investigate its negative impact on weakly-supervised audio-visual learning. To address these issues, we propose a modality-aware contrastive instance learning with self-distillation (MACIL-SD) strategy. Specifically, we leverage a lightweight two-stream network to generate audio and visual bags, in which unimodal background, violent, and normal instances are clustered into semi-bags in an unsupervised way. Then audio and visual violent semi-bag representations are assembled as positive pairs, and violent semi-bags are combined with background and normal instances in the opposite modality as contrastive negative pairs. Furthermore, a self-distillation module is applied to transfer unimodal visual knowledge to the audio-visual model, which alleviates noises and closes the semantic gap between unimodal and multimodal features. Experiments show that our framework outperforms previous methods with lower complexity on the large-scale XD-Violence dataset. Results also demonstrate that our proposed approach can be used as plug-in modules to enhance other networks. Codes are available at https://github.com/JustinYuu/MACIL_SD.

Jilan Xu, Junlin Hou, Yuejie Zhang et al.

In this paper, we consider the problem of open-vocabulary semantic segmentation (OVS), which aims to segment objects of arbitrary classes instead of pre-defined, closed-set categories. The main contributions are as follows: First, we propose a transformer-based model for OVS, termed as OVSegmentor, which only exploits web-crawled image-text pairs for pre-training without using any mask annotations. OVSegmentor assembles the image pixels into a set of learnable group tokens via a slot-attention based binding module, and aligns the group tokens to the corresponding caption embedding. Second, we propose two proxy tasks for training, namely masked entity completion and cross-image mask consistency. The former aims to infer all masked entities in the caption given the group tokens, that enables the model to learn fine-grained alignment between visual groups and text entities. The latter enforces consistent mask predictions between images that contain shared entities, which encourages the model to learn visual invariance. Third, we construct CC4M dataset for pre-training by filtering CC12M with frequently appeared entities, which significantly improves training efficiency. Fourth, we perform zero-shot transfer on three benchmark datasets, PASCAL VOC 2012, PASCAL Context, and COCO Object. Our model achieves superior segmentation results over the state-of-the-art method by using only 3\% data (4M vs 134M) for pre-training. Code and pre-trained models will be released for future research.

Junlin Hou, Fan Xiao, Jilan Xu et al.

The ultra-wide optical coherence tomography angiography (OCTA) has become an important imaging modality in diabetic retinopathy (DR) diagnosis. However, there are few researches focusing on automatic DR analysis using ultra-wide OCTA. In this paper, we present novel and practical deep-learning solutions based on ultra-wide OCTA for the Diabetic Retinopathy Analysis Challenge (DRAC). In the segmentation of DR lesions task, we utilize UNet and UNet++ to segment three lesions with strong data augmentation and model ensemble. In the image quality assessment task, we create an ensemble of InceptionV3, SE-ResNeXt, and Vision Transformer models. Pre-training on the large dataset as well as the hybrid MixUp and CutMix strategy are both adopted to boost the generalization ability of our model. In the DR grading task, we build a Vision Transformer (ViT) and fnd that the ViT model pre-trained on color fundus images serves as a useful substrate for OCTA images. Our proposed methods ranked 4th, 3rd, and 5th on the three leaderboards of DRAC, respectively. The source code will be made available at https://github.com/FDU-VTS/DRAC.

Xinyu Huang, Yi-Jie Huang, Youcai Zhang et al.

In this paper, we introduce the Recognize Anything Plus Model (RAM++), an open-set image tagging model effectively leveraging multi-grained text supervision. Previous approaches (e.g., CLIP) primarily utilize global text supervision paired with images, leading to sub-optimal performance in recognizing multiple individual semantic tags. In contrast, RAM++ seamlessly integrates individual tag supervision with global text supervision, all within a unified alignment framework. This integration not only ensures efficient recognition of predefined tag categories, but also enhances generalization capabilities for diverse open-set categories. Furthermore, RAM++ employs large language models (LLMs) to convert semantically constrained tag supervision into more expansive tag description supervision, thereby enriching the scope of open-set visual description concepts. Comprehensive evaluations on various image recognition benchmarks demonstrate RAM++ exceeds existing state-of-the-art (SOTA) open-set image tagging models on most aspects. Specifically, for predefined commonly used tag categories, RAM++ showcases 10.2 mAP and 15.4 mAP enhancements over CLIP on OpenImages and ImageNet. For open-set categories beyond predefined, RAM++ records improvements of 5.0 mAP and 6.4 mAP over CLIP and RAM respectively on OpenImages. For diverse human-object interaction phrases, RAM++ achieves 7.8 mAP and 4.7 mAP improvements on the HICO benchmark. Code, datasets and pre-trained models are available at \url{https://github.com/xinyu1205/recognize-anything}.

Rongzhi Zhang, Rui Feng, Zhihan Zhang et al.

Rubric-based RL is a promising route for extending reinforcement learning beyond verifiable rewards, yet existing methods optimize rubrics while treating the query distribution as fixed. We identify a structural bottleneck: rubric quality is constrained by query structure. Open-ended queries yield vague rubrics; naively narrowing them introduces fabricated references that no model can verify, so all responses fail and training receives no reward signal. We present QUBRIC, a framework that co-designs queries and rubrics. Teacher-derived key points ground the rewriting of open-ended queries into scenario-based, evaluable questions. Contrastive rubric generation then turns teacher-policy gaps into query-level criteria, and learnability filtering retains only informative query-rubric pairs for GRPO training. QUBRIC achieves a +5.5 point gain on ArenaHard over the SFT baseline. Trained only on instruction-following data, it further transfers to three held-out benchmarks spanning legal, moral, and narrative reasoning (+6.3 points on average), with improvements concentrated in reasoning-related dimensions. These results provide evidence that co-designing queries and rubrics can make rubric-based RL a practical complement to RLVR beyond strictly verifiable tasks.

Chenghao Yin, Da Huang, Di Yang et al.

The development of robust and generalizable robot learning models is critically contingent upon the availability of large-scale, diverse training data and reliable evaluation benchmarks. Collecting data in the physical world poses prohibitive costs and scalability challenges, and prevailing simulation benchmarks frequently suffer from fragmentation, narrow scope, or insufficient fidelity to enable effective sim-to-real transfer. To address these challenges, we introduce Genie Sim 3.0, a unified simulation platform for robotic manipulation. We present Genie Sim Generator, a large language model (LLM)-powered tool that constructs high-fidelity scenes from natural language instructions. Its principal strength resides in rapid and multi-dimensional generalization, facilitating the synthesis of diverse environments to support scalable data collection and robust policy evaluation. We introduce the first benchmark that pioneers the application of LLM for automated evaluation. It leverages LLM to mass-generate evaluation scenarios and employs Vision-Language Model (VLM) to establish an automated assessment pipeline. We also release an open-source dataset comprising more than 10,000 hours of synthetic data across over 200 tasks. Through systematic experimentation, we validate the robust zero-shot sim-to-real transfer capability of our open-source dataset, demonstrating that synthetic data can server as an effective substitute for real-world data under controlled conditions for scalable policy training. For code and dataset details, please refer to: https://github.com/AgibotTech/genie_sim.

Rui Feng, Qi Zhu, Huan Tran et al.

Recent works have shown the promise of learning pre-trained models for 3D molecular representation. However, existing pre-training models focus predominantly on equilibrium data and largely overlook off-equilibrium conformations. It is challenging to extend these methods to off-equilibrium data because their training objective relies on assumptions of conformations being the local energy minima. We address this gap by proposing a force-centric pretraining model for 3D molecular conformations covering both equilibrium and off-equilibrium data. For off-equilibrium data, our model learns directly from their atomic forces. For equilibrium data, we introduce zero-force regularization and forced-based denoising techniques to approximate near-equilibrium forces. We obtain a unified pre-trained model for 3D molecular representation with over 15 million diverse conformations. Experiments show that, with our pre-training objective, we increase forces accuracy by around 3 times compared to the un-pre-trained Equivariant Transformer model. By incorporating regularizations on equilibrium data, we solved the problem of unstable MD simulations in vanilla Equivariant Transformers, achieving state-of-the-art simulation performance with 2.45 times faster inference time than NequIP. As a powerful molecular encoder, our pre-trained model achieves on-par performance with state-of-the-art property prediction tasks.

Xinyu Huang, Youcai Zhang, Ying Cheng et al.

Vision-Language Pre-training (VLP) with large-scale image-text pairs has demonstrated superior performance in various fields. However, the image-text pairs co-occurrent on the Internet typically lack explicit alignment information, which is suboptimal for VLP. Existing methods proposed to adopt an off-the-shelf object detector to utilize additional image tag information. However, the object detector is time-consuming and can only identify the pre-defined object categories, limiting the model capacity. Inspired by the observation that the texts incorporate incomplete fine-grained image information, we introduce IDEA, which stands for increasing text diversity via online multi-label recognition for VLP. IDEA shows that multi-label learning with image tags extracted from the texts can be jointly optimized during VLP. Moreover, IDEA can identify valuable image tags online to provide more explicit textual supervision. Comprehensive experiments demonstrate that IDEA can significantly boost the performance on multiple downstream datasets with a small extra computational cost.

Shuaicheng Li, Feng Zhang, Rui-Wei Zhao et al.

It has been found that temporal action proposal generation, which aims to discover the temporal action instances within the range of the start and end frames in the untrimmed videos, can largely benefit from proper temporal and semantic context exploitation. The latest efforts were dedicated to considering the temporal context and similarity-based semantic contexts through self-attention modules. However, they still suffer from cluttered background information and limited contextual feature learning. In this paper, we propose a novel Pyramid Region-based Slot Attention (PRSlot) module to address these issues. Instead of using the similarity computation, our PRSlot module directly learns the local relations in an encoder-decoder manner and generates the representation of a local region enhanced based on the attention over input features called \textit{slot}. Specifically, upon the input snippet-level features, PRSlot module takes the target snippet as \textit{query}, its surrounding region as \textit{key} and then generates slot representations for each \textit{query-key} slot by aggregating the local snippet context with a parallel pyramid strategy. Based on PRSlot modules, we present a novel Pyramid Region-based Slot Attention Network termed PRSA-Net to learn a unified visual representation with rich temporal and semantic context for better proposal generation. Extensive experiments are conducted on two widely adopted THUMOS14 and ActivityNet-1.3 benchmarks. Our PRSA-Net outperforms other state-of-the-art methods. In particular, we improve the AR@100 from the previous best 50.67% to 56.12% for proposal generation and raise the mAP under 0.5 tIoU from 51.9\% to 58.7\% for action detection on THUMOS14. \textit{Code is available at} \url{https://github.com/handhand123/PRSA-Net}

Lingkai Kong, Jiaming Cui, Yuchen Zhuang et al.

Decision-focused learning (DFL) was recently proposed for stochastic optimization problems that involve unknown parameters. By integrating predictive modeling with an implicitly differentiable optimization layer, DFL has shown superior performance to the standard two-stage predict-then-optimize pipeline. However, most existing DFL methods are only applicable to convex problems or a subset of nonconvex problems that can be easily relaxed to convex ones. Further, they can be inefficient in training due to the requirement of solving and differentiating through the optimization problem in every training iteration. We propose SO-EBM, a general and efficient DFL method for stochastic optimization using energy-based models. Instead of relying on KKT conditions to induce an implicit optimization layer, SO-EBM explicitly parameterizes the original optimization problem using a differentiable optimization layer based on energy functions. To better approximate the optimization landscape, we propose a coupled training objective that uses a maximum likelihood loss to capture the optimum location and a distribution-based regularizer to capture the overall energy landscape. Finally, we propose an efficient training procedure for SO-EBM with a self-normalized importance sampler based on a Gaussian mixture proposal. We evaluate SO-EBM in three applications: power scheduling, COVID-19 resource allocation, and non-convex adversarial security game, demonstrating the effectiveness and efficiency of SO-EBM.

Junlin Hou, Jilan Xu, Nan Zhang et al.

This paper presents our solution for the 2nd COVID-19 Severity Detection Competition. This task aims to distinguish the Mild, Moderate, Severe, and Critical grades in COVID-19 chest CT images. In our approach, we devise a novel infection-aware 3D Contrastive Mixup Classification network for severity grading. Specifcally, we train two segmentation networks to first extract the lung region and then the inner lesion region. The lesion segmentation mask serves as complementary information for the original CT slices. To relieve the issue of imbalanced data distribution, we further improve the advanced Contrastive Mixup Classification network by weighted cross-entropy loss. On the COVID-19 severity detection leaderboard, our approach won the first place with a Macro F1 Score of 51.76%. It significantly outperforms the baseline method by over 11.46%.

Junlin Hou, Jilan Xu, Nan Zhang et al.

This paper presents our solution for the 2nd COVID-19 Competition, occurring in the framework of the AIMIA Workshop at the European Conference on Computer Vision (ECCV 2022). In our approach, we employ the winning solution last year which uses a strong 3D Contrastive Mixup Classifcation network (CMC v1) as the baseline method, composed of contrastive representation learning and mixup classification. In this paper, we propose CMC v2 by introducing natural video priors to COVID-19 diagnosis. Specifcally, we adapt a pre-trained (on video dataset) video transformer backbone to COVID-19 detection. Moreover, advanced training strategies, including hybrid mixup and cutmix, slicelevel augmentation, and small resolution training are also utilized to boost the robustness and the generalization ability of the model. Among 14 participating teams, CMC v2 ranked 1st in the 2nd COVID-19 Competition with an average Macro F1 Score of 89.11%.

Jiashuo Yu, Junfu Pu, Ying Cheng et al.

Although audio-visual representation has been proved to be applicable in many downstream tasks, the representation of dancing videos, which is more specific and always accompanied by music with complex auditory contents, remains challenging and uninvestigated. Considering the intrinsic alignment between the cadent movement of dancer and music rhythm, we introduce MuDaR, a novel Music-Dance Representation learning framework to perform the synchronization of music and dance rhythms both in explicit and implicit ways. Specifically, we derive the dance rhythms based on visual appearance and motion cues inspired by the music rhythm analysis. Then the visual rhythms are temporally aligned with the music counterparts, which are extracted by the amplitude of sound intensity. Meanwhile, we exploit the implicit coherence of rhythms implied in audio and visual streams by contrastive learning. The model learns the joint embedding by predicting the temporal consistency between audio-visual pairs. The music-dance representation, together with the capability of detecting audio and visual rhythms, can further be applied to three downstream tasks: (a) dance classification, (b) music-dance retrieval, and (c) music-dance retargeting. Extensive experiments demonstrate that our proposed framework outperforms other self-supervised methods by a large margin.

Bo Qian, Hao Chen, Xiangning Wang et al.

Computer-assisted automatic analysis of diabetic retinopathy (DR) is of great importance in reducing the risks of vision loss and even blindness. Ultra-wide optical coherence tomography angiography (UW-OCTA) is a non-invasive and safe imaging modality in DR diagnosis system, but there is a lack of publicly available benchmarks for model development and evaluation. To promote further research and scientific benchmarking for diabetic retinopathy analysis using UW-OCTA images, we organized a challenge named "DRAC - Diabetic Retinopathy Analysis Challenge" in conjunction with the 25th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2022). The challenge consists of three tasks: segmentation of DR lesions, image quality assessment and DR grading. The scientific community responded positively to the challenge, with 11, 12, and 13 teams from geographically diverse institutes submitting different solutions in these three tasks, respectively. This paper presents a summary and analysis of the top-performing solutions and results for each task of the challenge. The obtained results from top algorithms indicate the importance of data augmentation, model architecture and ensemble of networks in improving the performance of deep learning models. These findings have the potential to enable new developments in diabetic retinopathy analysis. The challenge remains open for post-challenge registrations and submissions for benchmarking future methodology developments.

Junlin Hou, Jilan Xu, Rui Feng et al.

This paper presents our solution for the 2nd COVID-19 Competition, occurring in the framework of the AIMIA Workshop in the European Conference on Computer Vision (ECCV 2022). In our approach, we employ an effective 3D Contrastive Mixup Classification network for COVID-19 diagnosis on chest CT images, which is composed of contrastive representation learning and mixup classification. For the COVID-19 detection challenge, our approach reaches 0.9245 macro F1 score on 484 validation CT scans, which significantly outperforms the baseline method by 16.5%. In the COVID-19 severity detection challenge, our approach achieves 0.7186 macro F1 score on 61 validation samples, which also surpasses the baseline by 8.86%.

Jinyu Liu, Ying Cheng, Yuejie Zhang et al.

Visual-only self-supervised learning has achieved significant improvement in video representation learning. Existing related methods encourage models to learn video representations by utilizing contrastive learning or designing specific pretext tasks. However, some models are likely to focus on the background, which is unimportant for learning video representations. To alleviate this problem, we propose a new view called long-range residual frame to obtain more motion-specific information. Based on this, we propose the Motion-Contrastive Perception Network (MCPNet), which consists of two branches, namely, Motion Information Perception (MIP) and Contrastive Instance Perception (CIP), to learn generic video representations by focusing on the changing areas in videos. Specifically, the MIP branch aims to learn fine-grained motion features, and the CIP branch performs contrastive learning to learn overall semantics information for each instance. Experiments on two benchmark datasets UCF-101 and HMDB-51 show that our method outperforms current state-of-the-art visual-only self-supervised approaches.

Yize Cai, Rui Feng, Anlan Yu et al.

Human Activity Recognition (HAR) from wearable sensors supports broad healthcare and behavior science applications. However, data heterogeneity and the scarcity of labeled data limit its real-world generalization. Recent advances in self-supervised learning (SSL) in vision and language domains have shown strong capability for learning generalizable representations from unlabeled data. Yet, few studies have systematically compared the generalization performance of SSL methods or explored how to adapt them for generalizable HAR. To address these gaps, we present BenchHAR, a unified framework for evaluating the generalization capability of SSL methods for sensor-based HAR on unseen target distributions. BenchHAR curates a large-scale dataset (~258K samples) and evaluates eight representative SSL methods across 12 encoder-classifier architectures. Our results reveal that existing SSL methods struggle to achieve satisfactory generalization performance. We find that: (1) For HAR models, the hybrid paradigm (combining reconstruction and contrastive pretraining) achieves the best overall performance. The CNN encoder exhibits the strongest ability to learn generalizable representations, while more expressive classifier architectures further improve generalization. (2) For data scale, increasing the amount of pretraining data from downstream activity classes consistently improves generalization, while adding more labeled data yields limited gains. Interestingly, incorporating unlabeled data from non-downstream activity classes does not improve generalization. (3) Sensor data collected from custom-grade devices generalizes better than that from research-grade devices, and data from limb transfers more effectively to trunk positions. BenchHAR provides a unified benchmark and actionable insights for generalizable sensor-based HAR systems. Our code is available at https://github.com/saiketa/HAR-Bench.

Qingqiu Li, Jilan Xu, Runtian Yuan et al.

Automatic generation of radiology reports holds crucial clinical value, as it can alleviate substantial workload on radiologists and remind less experienced ones of potential anomalies. Despite the remarkable performance of various image captioning methods in the natural image field, generating accurate reports for medical images still faces challenges, i.e., disparities in visual and textual data, and lack of accurate domain knowledge. To address these issues, we propose an enhanced knowledge injection framework, which utilizes two branches to extract different types of knowledge. The Weighted Concept Knowledge (WCK) branch is responsible for introducing clinical medical concepts weighted by TF-IDF scores. The Multimodal Retrieval Knowledge (MRK) branch extracts triplets from similar reports, emphasizing crucial clinical information related to entity positions and existence. By integrating this finer-grained and well-structured knowledge with the current image, we are able to leverage the multi-source knowledge gain to ultimately facilitate more accurate report generation. Extensive experiments have been conducted on two public benchmarks, demonstrating that our method achieves superior performance over other state-of-the-art methods. Ablation studies further validate the effectiveness of two extracted knowledge sources.

Weiwei Tian, Xinyu Huang, Tianhao Cheng et al.

Pediatric pneumonia is the leading cause of death among children under five years worldwide, imposing a substantial burden on affected families. Currently, there are three significant hurdles in diagnosing and treating pediatric pneumonia. Firstly, pediatric pneumonia shares similar symptoms with other respiratory diseases, making rapid and accurate differential diagnosis challenging. Secondly, primary hospitals often lack sufficient medical resources and experienced doctors. Lastly, providing personalized diagnostic reports and treatment recommendations is labor-intensive and time-consuming. To tackle these challenges, we proposed a Medical Multimodal Large Language Model for Pediatric Pneumonia (P2Med-MLLM). It was capable of handling diverse clinical tasks, such as generating free-text radiology reports and medical records within a unified framework. Specifically, P2Med-MLLM can process both pure text and image-text data, trained on an extensive and large-scale dataset (P2Med-MD), including real clinical information from 163,999 outpatient and 8,684 inpatient cases. This dataset comprised 2D chest X-ray images, 3D chest CT images, corresponding radiology reports, and outpatient and inpatient records. We designed a three-stage training strategy to enable P2Med-MLLM to comprehend medical knowledge and follow instructions for various clinical tasks. To rigorously evaluate P2Med-MLLM's performance, we developed P2Med-MBench, a benchmark consisting of 642 meticulously verified samples by pediatric pulmonology specialists, covering six clinical decision-support tasks and a balanced variety of diseases. The automated scoring results demonstrated the superiority of P2Med-MLLM. This work plays a crucial role in assisting primary care doctors with prompt disease diagnosis and treatment planning, reducing severe symptom mortality rates, and optimizing the allocation of medical resources.

Xinyu Huang, Yuhao Dong, Weiwei Tian et al.

State-of-the-art large multi-modal models (LMMs) face challenges when processing high-resolution images, as these inputs are converted into enormous visual tokens, many of which are irrelevant to the downstream task. In this paper, we propose Multi-turn Grounding-based Policy Optimization (MGPO), an end-to-end reinforcement learning (RL) framework that enables LMMs to iteratively focus on key visual regions by automatically cropping sub-images, based on model-predicted grounding coordinates within a multi-turn conversation framework. Compared to supervised fine-tuning (SFT), which requires costly additional grounding annotations, our approach highlights that LMMs can emerge robust grounding abilities during the RL training process, leveraging only a binary reward function derived from the correctness of the final answer. Additionally, we observe that LMMs struggle to autonomously trigger visual grounding during the rollout process. To address this cold start problem, we design a multi-turn conversational template and restrict policy loss computation to model outputs generated across multiple dialogue rounds, thereby promoting stable optimization. Extensive experiments demonstrate that, when trained on standard visual-question-short answering data without grounding annotations, MGPO effectively elicits stronger grounding capabilities compared to GRPO, leading to 5.4\% improvement on in-distribution MME-Realworld and 5.2\% improvement on the challenging out-of-distribution (OOD) V* Bench. Notably, MGPO post-training on Qwen2.5-VL-7B with 21K samples surpasses OpenAI's o1 and GPT-4o models on the OOD V* Bench. Codes are available at https://github.com/EvolvingLMMs-Lab/MGPO.

Hongzhou Yu, Tianhao Cheng, Yingwen Wang et al.

Recent advancements in large language models (LLMs) have shown promise in medical applications such as disease diagnosis and treatment planning. However, most existing medical LLMs struggle with the deep reasoning required for complex medical problems, such as differential diagnosis and medication recommendations. We propose FineMedLM-o1, which leverages high-quality medical synthetic data and long-form reasoning data for Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO), enabling advanced dialogue and deep reasoning capabilities. Additionally, we introduce Test-Time Training (TTT) in the medical domain for the first time, facilitating domain adaptation and ensuring reliable, accurate reasoning. Experimental results demonstrate that FineMedLM-o1 achieves a 23% average performance improvement over prior models on key medical benchmarks. Furthermore, the introduction of TTT provides an additional 14% performance boost, highlighting its effectiveness in enhancing medical reasoning capabilities. To support this process, we also propose a novel method for synthesizing medical dialogue. Compared to other open-source datasets, our dataset stands out as superior in both quality and complexity. The project and data will be released on GitHub.

Yin-Long Liu, Rui Feng, Jia-Hong Yuan et al.

Compared to other clinical screening techniques, speech-and-language-based automated Alzheimer's disease (AD) detection methods are characterized by their non-invasiveness, cost-effectiveness, and convenience. Previous studies have demonstrated the efficacy of fine-tuning pre-trained language models (PLMs) for AD detection. However, the objective of this traditional fine-tuning method, which involves inputting only transcripts, is inconsistent with the masked language modeling (MLM) task used during the pre-training phase of PLMs. In this paper, we investigate prompt-based fine-tuning of PLMs, converting the classification task into a MLM task by inserting prompt templates into the transcript inputs. We also explore the impact of incorporating pause information from forced alignment into manual transcripts. Additionally, we compare the performance of various automatic speech recognition (ASR) models and select the Whisper model to generate ASR-based transcripts for comparison with manual transcripts. Furthermore, majority voting and ensemble techniques are applied across different PLMs (BERT and RoBERTa) using different random seeds. Ultimately, we obtain maximum detection accuracy of 95.8% (with mean 87.9%, std 3.3%) using manual transcripts, achieving state-of-the-art performance for AD detection using only transcripts on the ADReSS test set.

Jianfa Bai, Kejin Lu, Runtian Yuan et al.

Robust detection of COVID-19 from chest CT remains challenging in multi-institutional settings due to substantial source shift, source imbalance, and hidden test-source identities. In this work, we propose a three-stage source-aware multi-expert framework for multi-source COVID-19 CT classification. First, we build a lung-aware 3D expert by combining original CT volumes and lung-extracted CT volumes for volumetric classification. Second, we develop two MedSigLIP-based experts: a slice-wise representation and probability learning module, and a Transformer-based inter-slice context modeling module for capturing cross-slice dependency. Third, we train a source classifier to predict the latent source identity of each test scan. By leveraging the predicted source information, we perform model fusion and voting based on different experts. On the validation set covering all four sources, the Stage 1 model achieves the best macro-F1 of 0.9711, ACC of 0.9712, and AUC of 0.9791. Stage~2a and Stage~2b achieve the best AUC scores of 0.9864 and 0.9854, respectively. Stage~3 source classifier reaches 0.9107 ACC and 0.9114 F1. These results demonstrate that source-aware expert modeling and hierarchical voting provide an effective solution for robust COVID-19 CT classification under heterogeneous multi-source conditions.

Anyang Song, Ying Cheng, Yiqian Xu et al.

Large Language Models (LLMs) alignment is constantly evolving. Machine-Generated Text (MGT) is becoming increasingly difficult to distinguish from Human-Written Text (HWT). This has exacerbated abuse issues such as fake news and online fraud. Fine-tuned detectors' generalization ability is highly dependent on dataset quality, and simply expanding the sources of MGT is insufficient. Further augment of generation process is required. According to HC-Var's theory, enhancing the alignment of generated text can not only facilitate attacks on existing detectors to test their robustness, but also help improve the generalization ability of detectors fine-tuned on it. Therefore, we propose \textbf{M}achine-\textbf{A}ugment-\textbf{G}enerated Text via \textbf{A}lignment (MAGA). MAGA's pipeline achieves comprehensive alignment from prompt construction to reasoning process, among which \textbf{R}einforced \textbf{L}earning from \textbf{D}etectors \textbf{F}eedback (RLDF), systematically proposed by us, serves as a key component. In our experiments, the RoBERTa detector fine-tuned on MAGA training set achieved an average improvement of 4.60\% in generalization detection AUC. MAGA Dataset caused an average decrease of 8.13\% in the AUC of the selected detectors, expecting to provide indicative significance for future research on the generalization detection ability of detectors.

Kejin Lu, Jianfa Bai, Qingqiu Li et al.

Accurate classification of lung diseases from chest CT scans plays an important role in computer-aided diagnosis systems. However, medical imaging datasets often suffer from severe class imbalance, which may significantly degrade the performance of deep learning models, especially for minority disease categories. To address this issue, we propose a gender-aware two-stage lung disease classification framework. The proposed approach explicitly incorporates gender information into the disease recognition pipeline. In the first stage, a gender classifier is trained to predict the patient's gender from CT scans. In the second stage, the input CT image is routed to a corresponding gender-specific disease classifier to perform final disease prediction. This design enables the model to better capture gender-related imaging characteristics and alleviate the influence of imbalanced data distribution. Experimental results demonstrate that the proposed method improves the recognition performance for minority disease categories, particularly squamous cell carcinoma, while maintaining competitive performance on other classes.

Fan Xiao, Junlin Hou, Ruiwei Zhao et al.

Diabetic retinopathy (DR) is a leading cause of blindness worldwide and a common complication of diabetes. As two different imaging tools for DR grading, color fundus photography (CFP) and infrared fundus photography (IFP) are highly-correlated and complementary in clinical applications. To the best of our knowledge, this is the first study that explores a novel multi-modal deep learning framework to fuse the information from CFP and IFP towards more accurate DR grading. Specifically, we construct a dual-stream architecture Cross-Fundus Transformer (CFT) to fuse the ViT-based features of two fundus image modalities. In particular, a meticulously engineered Cross-Fundus Attention (CFA) module is introduced to capture the correspondence between CFP and IFP images. Moreover, we adopt both the single-modality and multi-modality supervisions to maximize the overall performance for DR grading. Extensive experiments on a clinical dataset consisting of 1,713 pairs of multi-modal fundus images demonstrate the superiority of our proposed method. Our code will be released for public access.

Liuyu Wu, Rui Feng, Jie Li et al.

The automated evaluation of cognitive status utilizing multimedia technologies presents a promising frontier in early dementia diagnosis. However, the development of robust machine learning models for cognitive impairment detection is frequently hindered by the scarcity of large-scale, strictly synchronized, and clinically validated multimodal datasets. To bridge this critical gap, we introduce the CogPic database, a comprehensive multimodal benchmark meticulously designed for fine-grained cognitive impairment detection. The dataset comprises strictly synchronized audio, visual, and linguistic data continuously collected from 574 participants during a naturalistic picture description task. To establish highly reliable diagnostic ground truth, expert clinical neuropsychologists conducted exhaustive evaluations, stratifying participants into distinct cognitive groups through a comprehensive clinical consensus. Consequently, CogPic stands as the largest, most modality-rich, and most meticulously evaluated dataset of its kind to date. By conducting extensive benchmark experiments on the CogPic dataset, we establish an exceptionally robust, unbiased, and clinically generalizable foundation to propel future multimedia research in automated cognitive health assessment. Detailed information and access application procedures for our CogPic database are available at https://cogpic.github.io/.

Ling Zhou, Runtian Yuan, Yi Liu et al.

Ultrasound imaging is a prevalent diagnostic tool known for its simplicity and non-invasiveness. However, its inherent characteristics often introduce substantial noise, posing considerable challenges for automated lesion or organ segmentation in ultrasound video sequences. To address these limitations, we propose the Dual Semantic-Aware Network (DSANet), a novel framework designed to enhance noise robustness in ultrasound video segmentation by fostering mutual semantic awareness between local and global features. Specifically, we introduce an Adjacent-Frame Semantic-Aware (AFSA) module, which constructs a channel-wise similarity matrix to guide feature fusion across adjacent frames, effectively mitigating the impact of random noise without relying on pixel-level relationships. Additionally, we propose a Local-and-Global Semantic-Aware (LGSA) module that reorganizes and fuses temporal unconditional local features, which capture spatial details independently at each frame, with conditional global features that incorporate temporal context from adjacent frames. This integration facilitates multi-level semantic representation, significantly improving the model's resilience to noise interference. Extensive evaluations on four benchmark datasets demonstrate that DSANet substantially outperforms state-of-the-art methods in segmentation accuracy. Moreover, since our model avoids pixel-level feature dependencies, it achieves significantly higher inference FPS than video-based methods, and even surpasses some image-based models. Code can be found in \href{https://github.com/ZhouL2001/DSANet}{DSANet}

Jilan Xu, Yifei Huang, Junlin Hou et al.

Understanding human actions from videos of first-person view poses significant challenges. Most prior approaches explore representation learning on egocentric videos only, while overlooking the potential benefit of exploiting existing large-scale third-person videos. In this paper, (1) we develop EgoInstructor, a retrieval-augmented multimodal captioning model that automatically retrieves semantically relevant third-person instructional videos to enhance the video captioning of egocentric videos. (2) For training the cross-view retrieval module, we devise an automatic pipeline to discover ego-exo video pairs from distinct large-scale egocentric and exocentric datasets. (3) We train the cross-view retrieval module with a novel EgoExoNCE loss that pulls egocentric and exocentric video features closer by aligning them to shared text features that describe similar actions. (4) Through extensive experiments, our cross-view retrieval module demonstrates superior performance across seven benchmarks. Regarding egocentric video captioning, EgoInstructor exhibits significant improvements by leveraging third-person videos as references. Project page is available at: https://jazzcharles.github.io/Egoinstructor/

Bowen Zhao, Changkai Ji, Yuejie Zhang et al.

With the Generative Pre-trained Transformer 3.5 (GPT-3.5) exhibiting remarkable reasoning and comprehension abilities in Natural Language Processing (NLP), most Question Answering (QA) research has primarily centered around general QA tasks based on GPT, neglecting the specific challenges posed by Complex Table QA. In this paper, we propose to incorporate GPT-3.5 to address such challenges, in which complex tables are reconstructed into tuples and specific prompt designs are employed for dialogues. Specifically, we encode each cell's hierarchical structure, position information, and content as a tuple. By enhancing the prompt template with an explanatory description of the meaning of each tuple and the logical reasoning process of the task, we effectively improve the hierarchical structure awareness capability of GPT-3.5 to better parse the complex tables. Extensive experiments and results on Complex Table QA datasets, i.e., the open-domain dataset HiTAB and the aviation domain dataset AIT-QA show that our approach significantly outperforms previous work on both datasets, leading to state-of-the-art (SOTA) performance.

Jilan Xu, Yifei Huang, Baoqi Pei et al.

Generating videos in the first-person perspective has broad application prospects in the field of augmented reality and embodied intelligence. In this work, we explore the cross-view video prediction task, where given an exo-centric video, the first frame of the corresponding ego-centric video, and textual instructions, the goal is to generate futur frames of the ego-centric video. Inspired by the notion that hand-object interactions (HOI) in ego-centric videos represent the primary intentions and actions of the current actor, we present EgoExo-Gen that explicitly models the hand-object dynamics for cross-view video prediction. EgoExo-Gen consists of two stages. First, we design a cross-view HOI mask prediction model that anticipates the HOI masks in future ego-frames by modeling the spatio-temporal ego-exo correspondence. Next, we employ a video diffusion model to predict future ego-frames using the first ego-frame and textual instructions, while incorporating the HOI masks as structural guidance to enhance prediction quality. To facilitate training, we develop an automated pipeline to generate pseudo HOI masks for both ego- and exo-videos by exploiting vision foundation models. Extensive experiments demonstrate that our proposed EgoExo-Gen achieves better prediction performance compared to previous video prediction models on the Ego-Exo4D and H2O benchmark datasets, with the HOI masks significantly improving the generation of hands and interactive objects in the ego-centric videos.

Xiaze Zhang, Ziheng Ding, Qi Jing et al.

Point clouds have shown significant potential in various domains, including Simultaneous Localization and Mapping (SLAM). However, existing approaches either rely on dense point clouds to achieve high localization accuracy or use generalized descriptors to reduce map size. Unfortunately, these two aspects seem to conflict with each other. To address this limitation, we propose a unified architecture, DeepPointMap, achieving excellent preference on both aspects. We utilize neural network to extract highly representative and sparse neural descriptors from point clouds, enabling memory-efficient map representation and accurate multi-scale localization tasks (e.g., odometry and loop-closure). Moreover, we showcase the versatility of our framework by extending it to more challenging multi-agent collaborative SLAM. The promising results obtained in these scenarios further emphasize the effectiveness and potential of our approach.

Bowen Zhao, Tianhao Cheng, Yuejie Zhang et al.

Multimodal Question Answering (MMQA) is crucial as it enables comprehensive understanding and accurate responses by integrating insights from diverse data representations such as tables, charts, and text. Most existing researches in MMQA only focus on two modalities such as image-text QA, table-text QA and chart-text QA, and there remains a notable scarcity in studies that investigate the joint analysis of text, tables, and charts. In this paper, we present C$\text{T}^2$C-QA, a pioneering Chinese reasoning-based QA dataset that includes an extensive collection of text, tables, and charts, meticulously compiled from 200 selectively sourced webpages. Our dataset simulates real webpages and serves as a great test for the capability of the model to analyze and reason with multimodal data, because the answer to a question could appear in various modalities, or even potentially not exist at all. Additionally, we present AED (\textbf{A}llocating, \textbf{E}xpert and \textbf{D}esicion), a multi-agent system implemented through collaborative deployment, information interaction, and collective decision-making among different agents. Specifically, the Assignment Agent is in charge of selecting and activating expert agents, including those proficient in text, tables, and charts. The Decision Agent bears the responsibility of delivering the final verdict, drawing upon the analytical insights provided by these expert agents. We execute a comprehensive analysis, comparing AED with various state-of-the-art models in MMQA, including GPT-4. The experimental outcomes demonstrate that current methodologies, including GPT-4, are yet to meet the benchmarks set by our dataset.

Junlin Hou, Jilan Xu, Rui Feng et al.

Due to the complexity of medical image acquisition and the difficulty of annotation, medical image datasets inevitably contain noise. Noisy data with wrong labels affects the robustness and generalization ability of deep neural networks. Previous noise learning methods mainly considered noise arising from images being mislabeled, i.e. label noise, assuming that all mislabeled images are of high image quality. However, medical images are prone to suffering extreme quality issues, i.e. data noise, where discriminative visual features are missing for disease diagnosis. In this paper, we propose a noise learning framework, termed as QMix, that learns a robust disease diagnosis model under mixed noise. QMix alternates between sample separation and quality-aware semisupervised training in each training epoch. In the sample separation phase, we design a joint uncertainty-loss criterion to effectively separate (1) correctly labeled images; (2) mislabeled images with high quality and (3) mislabeled images with low quality. In the semi-supervised training phase, we train a disease diagnosis model to learn robust feature representation from the separated samples. Specifically, we devise a sample-reweighing loss to mitigate the effect of mislabeled images with low quality during training. Meanwhile, a contrastive enhancement loss is proposed to further distinguish mislabeled images with low quality from correctly labeled images. QMix achieved state-of-the-art disease diagnosis performance on five public retinal image datasets and exhibited substantial improvement on robustness against mixed noise.

Qingqiu Li, Runtian Yuan, Junlin Hou et al.

To make a more accurate diagnosis of COVID-19, we propose a straightforward yet effective model. Firstly, we analyse the characteristics of 3D CT scans and remove the non-lung parts, facilitating the model to focus on lesion-related areas and reducing computational cost. We use ResNeSt50 as the strong feature extractor, initializing it with pretrained weights which have COVID-19-specific prior knowledge. Our model achieves a Macro F1 Score of 0.94 on the validation set of the 4th COV19D Competition Challenge $\mathrm{I}$, surpassing the baseline by 16%. This indicates its effectiveness in distinguishing between COVID-19 and non-COVID-19 cases, making it a robust method for COVID-19 detection.

Qingqiu Li, Zihang Cui, Seongsu Bae et al.

Chest X-rays (CXRs) are the most frequently performed imaging examinations in clinical settings. Recent advancements in Large Multimodal Models (LMMs) have enabled automated CXR interpretation, enhancing diagnostic accuracy and efficiency. However, despite their strong visual understanding, current Medical LMMs (MLMMs) still face two major challenges: (1) Insufficient region-level understanding and interaction, and (2) Limited accuracy and interpretability due to single-step reasoning. In this paper, we empower MLMMs with anatomy-centric reasoning capabilities to enhance their interactivity and explainability. Specifically, we first propose an Anatomical Ontology-Guided Reasoning (AOR) framework, which centers on cross-modal region-level information to facilitate multi-step reasoning. Next, under the guidance of expert physicians, we develop AOR-Instruction, a large instruction dataset for MLMMs training. Our experiments demonstrate AOR's superior performance in both VQA and report generation tasks.

Runtian Yuan, Mohan Chen, Jilan Xu et al.

Referring Medical Image Sequence Segmentation (Ref-MISS) is a novel and challenging task that aims to segment anatomical structures in medical image sequences (\emph{e.g.} endoscopy, ultrasound, CT, and MRI) based on natural language descriptions. This task holds significant clinical potential and offers a user-friendly advancement in medical imaging interpretation. Existing 2D and 3D segmentation models struggle to explicitly track objects of interest across medical image sequences, and lack support for nteractive, text-driven guidance. To address these limitations, we propose Text-Promptable Propagation (TPP), a model designed for referring medical image sequence segmentation. TPP captures the intrinsic relationships among sequential images along with their associated textual descriptions. Specifically, it enables the recognition of referred objects through cross-modal referring interaction, and maintains continuous tracking across the sequence via Transformer-based triple propagation, using text embeddings as queries. To support this task, we curate a large-scale benchmark, Ref-MISS-Bench, which covers 4 imaging modalities and 20 different organs and lesions. Experimental results on this benchmark demonstrate that TPP consistently outperforms state-of-the-art methods in both medical segmentation and referring video object segmentation.

Runtian Yuan, Qingqiu Li, Junlin Hou et al.

In response to the need for rapid and accurate COVID-19 diagnosis during the global pandemic, we present a two-stage framework that leverages pseudo labels for domain adaptation to enhance the detection of COVID-19 from CT scans. By utilizing annotated data from one domain and non-annotated data from another, the model overcomes the challenge of data scarcity and variability, common in emergent health crises. The innovative approach of generating pseudo labels enables the model to iteratively refine its learning process, thereby improving its accuracy and adaptability across different hospitals and medical centres. Experimental results on COV19-CT-DB database showcase the model's potential to achieve high diagnostic precision, significantly contributing to efficient patient management and alleviating the strain on healthcare systems. Our method achieves 0.92 Macro F1 Score on the validation set of Covid-19 domain adaptation challenge.

Runtian Yuan, Qingqiu Li, Junlin Hou et al.

We present our solution for the Multi-Source COVID-19 Detection Challenge, which aims to classify chest CT scans into COVID and Non-COVID categories across data collected from four distinct hospitals and medical centers. A major challenge in this task lies in the domain shift caused by variations in imaging protocols, scanners, and patient populations across institutions. To enhance the cross-domain generalization of our model, we incorporate Variance Risk Extrapolation (VREx) into the training process. VREx encourages the model to maintain consistent performance across multiple source domains by explicitly minimizing the variance of empirical risks across environments. This regularization strategy reduces overfitting to center-specific features and promotes learning of domain-invariant representations. We further apply Mixup data augmentation to improve generalization and robustness. Mixup interpolates both the inputs and labels of randomly selected pairs of training samples, encouraging the model to behave linearly between examples and enhancing its resilience to noise and limited data. Our method achieves an average macro F1 score of 0.96 across the four sources on the validation set, demonstrating strong generalization.

Rui Feng, Yuang Chen, Yu Hu et al.

This letter introduces EGGCodec, a robust neural Encodec framework engineered for electroglottography (EGG) signal reconstruction and F0 extraction. We propose a multi-scale frequency-domain loss function to capture the nuanced relationship between original and reconstructed EGG signals, complemented by a time-domain correlation loss to improve generalization and accuracy. Unlike conventional Encodec models that extract F0 directly from features, EGGCodec leverages reconstructed EGG signals, which more closely correspond to F0. By removing the conventional GAN discriminator, we streamline EGGCodec's training process without compromising efficiency, incurring only negligible performance degradation. Trained on a widely used EGG-inclusive dataset, extensive evaluations demonstrate that EGGCodec outperforms state-of-the-art F0 extraction schemes, reducing mean absolute error (MAE) from 14.14 Hz to 13.69 Hz, and improving voicing decision error (VDE) by 38.2\%. Moreover, extensive ablation experiments validate the contribution of each component of EGGCodec.

Rui Feng, Zhiyao Luo, Wei Wang et al.

Automatic assessment of cognitive impairment from spontaneous speech offers a promising, non-invasive avenue for early cognitive screening. However, current approaches often lack generalizability when deployed across different languages and clinical settings, limiting their practical utility. In this study, we propose CogBench, the first benchmark designed to evaluate the cross-lingual and cross-site generalizability of large language models (LLMs) for speech-based cognitive impairment assessment. Using a unified multimodal pipeline, we evaluate model performance on three speech datasets spanning English and Mandarin: ADReSSo, NCMMSC2021-AD, and a newly collected test set, CIR-E. Our results show that conventional deep learning models degrade substantially when transferred across domains. In contrast, LLMs equipped with chain-of-thought prompting demonstrate better adaptability, though their performance remains sensitive to prompt design. Furthermore, we explore lightweight fine-tuning of LLMs via Low-Rank Adaptation (LoRA), which significantly improves generalization in target domains. These findings offer a critical step toward building clinically useful and linguistically robust speech-based cognitive assessment tools.

Liu He, Yuanchao Li, Rui Feng et al.

Gender bias has been widely observed in speech perception tasks, influenced by the fundamental voicing differences between genders. This study reveals a gender bias in the perception of Alzheimer's Disease (AD) speech. In a perception experiment involving 16 Chinese listeners evaluating both Chinese and Greek speech, we identified that male speech was more frequently identified as AD, with this bias being particularly pronounced in Chinese speech. Acoustic analysis showed that shimmer values in male speech were significantly associated with AD perception, while speech portion exhibited a significant negative correlation with AD identification. Although language did not have a significant impact on AD perception, our findings underscore the critical role of gender bias in AD speech perception. This work highlights the necessity of addressing gender bias when developing AD detection models and calls for further research to validate model performance across different linguistic contexts.

Qingqiu Li, Runtian Yuan, Junlin Hou et al.

To enable more accurate diagnosis of lung disease in chest CT scans, we propose a straightforward yet effective model. Firstly, we analyze the characteristics of 3D CT scans and remove non-lung regions, which helps the model focus on lesion-related areas and reduces computational cost. We adopt ResNeSt50 as a strong feature extractor, and use a weighted cross-entropy loss to mitigate class imbalance, especially for the underrepresented squamous cell carcinoma category. Our model achieves a Macro F1 Score of 0.80 on the validation set of the Fair Disease Diagnosis Challenge, demonstrating its strong performance in distinguishing between different lung conditions.

Anton Alyakin, Jaden Stryker, Daniel Alexander Alber et al.

General-purpose vision-language models (VLMs) demonstrate impressive capabilities, but their opaque training on uncurated internet data posse critical limitations for high-stakes decision-making, such as in neurosurgery. We present CNS-Obsidian, a neurosurgical VLM trained on peer-reviewed neurosurgical literature, and demonstrate its clinical utility compared with GPT-4o in a real-world setting. We compiled 23,984 articles from Neurosurgery Publications journals, yielding 78,853 figures and captions. Using GPT-4o and Claude Sonnet-3.5, we converted these image-text pairs into 263,064 training samples across three formats: instruction fine-tuning, multiple-choice questions, and differential diagnosis. We trained CNS-Obsidian, a fine-tune of the 34-billion parameter LLaVA-Next model. In a blinded, randomized deployment trial at NYU Langone Health (Aug 30-Nov 30, 2024), neurosurgeons were assigned to use either CNS-Obsidian or GPT-4o as a diagnostic co-pilot after patient consultations. Primary outcomes were diagnostic helpfulness and accuracy. CNS-Obsidian matched GPT-4o on synthetic questions (76.13% vs 77.54%, p=0.235), but only achieved 46.81% accuracy on human-generated questions versus GPT-4o's 65.70% (p<10-15). In the randomized trial, 70 consultations were evaluated (32 CNS-Obsidian, 38 GPT-4o) from 959 total consults. CNS-Obsidian received positive ratings in 40.62% of cases versus 57.89% for GPT-4o (p=0.230). Both models included correct diagnosis in approximately 60% of cases (59.38% vs 65.79%, p=0.626). Domain-specific VLMs trained on curated scientific literature can approach frontier model performance in specialized medical domains despite being orders of magnitude smaller and less expensive to train. However, low clinical utilization suggests chatbot interfaces may not align with specialist workflows, indicating need for alternative AI integration strategies.

Ruize Wang, Hui Xu, Ying Cheng et al.

Advertising platforms have evolved in estimating Lifetime Value (LTV) to better align with advertisers' true performance metric. However, the sparsity of real-world LTV data presents a significant challenge to LTV predictive model(i.e., pLTV), severely limiting the their capabilities. Therefore, we propose to utilize external data, in addition to the internal data of advertising platform, to expand the size of purchase samples and enhance the LTV prediction model of the advertising platform. To tackle the issue of data distribution shift between internal and external platforms, we introduce an Adaptive Difference Siamese Network (ADSNet), which employs cross-domain transfer learning to prevent negative transfer. Specifically, ADSNet is designed to learn information that is beneficial to the target domain. We introduce a gain evaluation strategy to calculate information gain, aiding the model in learning helpful information for the target domain and providing the ability to reject noisy samples, thus avoiding negative transfer. Additionally, we also design a Domain Adaptation Module as a bridge to connect different domains, reduce the distribution distance between them, and enhance the consistency of representation space distribution. We conduct extensive offline experiments and online A/B tests on a real advertising platform. Our proposed ADSNet method outperforms other methods, improving GINI by 2$\%$. The ablation study highlights the importance of the gain evaluation strategy in negative gain sample rejection and improving model performance. Additionally, ADSNet significantly improves long-tail prediction. The online A/B tests confirm ADSNet's efficacy, increasing online LTV by 3.47$\%$ and GMV by 3.89$\%$.

Qingqiu Li, Xiaohan Yan, Jilan Xu et al.

Learning medical visual representations through vision-language pre-training has reached remarkable progress. Despite the promising performance, it still faces challenges, i.e., local alignment lacks interpretability and clinical relevance, and the insufficient internal and external representation learning of image-report pairs. To address these issues, we propose an Anatomical Structure-Guided (ASG) framework. Specifically, we parse raw reports into triplets <anatomical region, finding, existence>, and fully utilize each element as supervision to enhance representation learning. For anatomical region, we design an automatic anatomical region-sentence alignment paradigm in collaboration with radiologists, considering them as the minimum semantic units to explore fine-grained local alignment. For finding and existence, we regard them as image tags, applying an image-tag recognition decoder to associate image features with their respective tags within each sample and constructing soft labels for contrastive learning to improve the semantic association of different image-report pairs. We evaluate the proposed ASG framework on two downstream tasks, including five public benchmarks. Experimental results demonstrate that our method outperforms the state-of-the-art methods.

Rui Feng, Huan Tran, Aubrey Toland et al.

Polymer simulation with both accuracy and efficiency is a challenging task. Machine learning (ML) forcefields have been developed to achieve both the accuracy of ab initio methods and the efficiency of empirical force fields. However, existing ML force fields are usually limited to single-molecule settings, and their simulations are not robust enough. In this paper, we present PolyGET, a new framework for Polymer Forcefields with Generalizable Equivariant Transformers. PolyGET is designed to capture complex quantum interactions between atoms and generalize across various polymer families, using a deep learning model called Equivariant Transformers. We propose a new training paradigm that focuses exclusively on optimizing forces, which is different from existing methods that jointly optimize forces and energy. This simple force-centric objective function avoids competing objectives between energy and forces, thereby allowing for learning a unified forcefield ML model over different polymer families. We evaluated PolyGET on a large-scale dataset of 24 distinct polymer types and demonstrated state-of-the-art performance in force accuracy and robust MD simulations. Furthermore, PolyGET can simulate large polymers with high fidelity to the reference ab initio DFT method while being able to generalize to unseen polymers.