Yitong Li

HyunJun Jung, Patrick Ruhkamp, Guangyao Zhai et al.

Learning-based methods to solve dense 3D vision problems typically train on 3D sensor data. The respectively used principle of measuring distances provides advantages and drawbacks. These are typically not compared nor discussed in the literature due to a lack of multi-modal datasets. Texture-less regions are problematic for structure from motion and stereo, reflective material poses issues for active sensing, and distances for translucent objects are intricate to measure with existing hardware. Training on inaccurate or corrupt data induces model bias and hampers generalisation capabilities. These effects remain unnoticed if the sensor measurement is considered as ground truth during the evaluation. This paper investigates the effect of sensor errors for the dense 3D vision tasks of depth estimation and reconstruction. We rigorously show the significant impact of sensor characteristics on the learned predictions and notice generalisation issues arising from various technologies in everyday household environments. For evaluation, we introduce a carefully designed dataset\footnote{dataset available at https://github.com/Junggy/HAMMER-dataset} comprising measurements from commodity sensors, namely D-ToF, I-ToF, passive/active stereo, and monocular RGB+P. Our study quantifies the considerable sensor noise impact and paves the way to improved dense vision estimates and targeted data fusion.

Tom Maye-Lasserre, Yitong Li, Bailiang Jian et al.

Modern 3D medical vision-language models (VLMs) can generate fluent radiology-style text while exhibit critically low pathology detection and output diversity, collapsing to generic templates that under-report rare yet critical findings. We identify this failure mode as Template Collapse. This failure stems from the unique constraints of 3D medical imaging, e.g., limited data, severe label imbalance, and weak signals from volumetric encoders. Under these constraints, text-generation objectives encourage shortcut learning and fluent but weakly grounded reports. We systematically diagnose the Template Collapse through clinical fidelity, output diversity, normal-template bias, and rare-finding survival. To mitigate it, we propose CLarGen, a decoupled framework that separates what to say (clinical detection) from how to say it (language synthesis). CLarGen uses (i) a Latent Query Transformer for multi-label pathology detection, (ii) pathology-guided retrieval for clinically matched exemplars, and (iii) a medical language model to synthesize the final report from detected findings and retrieved context. Across state-of-the-art 3D CT report generation baselines, CLarGen mitigates Template Collapse and substantially improves clinical accuracy (macro-F1 0.487 vs. 0.189; CRG 0.472 vs. 0.368) while maintaining fluent reporting. Our results suggest that explicit, measurable clinical grounding is essential for template-collapse-resistant 3D CT report generation. Code will be released upon acceptance.

Qi Zhu, Fei Mi, Zheng Zhang et al. · tsinghua

Incorporating external knowledge into the response generation process is essential to building more helpful and reliable dialog agents. However, collecting knowledge-grounded conversations is often costly, calling for a better pre-trained model for grounded dialog generation that generalizes well w.r.t. different types of knowledge. In this work, we propose KPT (Keyword-guided Pre-Training), a novel self-supervised pre-training method for grounded dialog generation without relying on extra knowledge annotation. Specifically, we use a pre-trained language model to extract the most uncertain tokens in the dialog as keywords. With these keywords, we construct two kinds of knowledge and pre-train a knowledge-grounded response generation model, aiming at handling two different scenarios: (1) the knowledge should be faithfully grounded; (2) it can be selectively used. For the former, the grounding knowledge consists of keywords extracted from the response. For the latter, the grounding knowledge is additionally augmented with keywords extracted from other utterances in the same dialog. Since the knowledge is extracted from the dialog itself, KPT can be easily performed on a large volume and variety of dialogue data. We considered three data sources (open-domain, task-oriented, conversational QA) with a total of 2.5M dialogues. We conduct extensive experiments on various few-shot knowledge-grounded generation tasks, including grounding on dialog acts, knowledge graphs, persona descriptions, and Wikipedia passages. Our comprehensive experiments and analyses demonstrate that KPT consistently outperforms state-of-the-art methods on these tasks with diverse grounding knowledge.

Xudong Han, Aili Shen, Yitong Li et al.

This paper presents fairlib, an open-source framework for assessing and improving classification fairness. It provides a systematic framework for quickly reproducing existing baseline models, developing new methods, evaluating models with different metrics, and visualizing their results. Its modularity and extensibility enable the framework to be used for diverse types of inputs, including natural language, images, and audio. In detail, we implement 14 debiasing methods, including pre-processing, at-training-time, and post-processing approaches. The built-in metrics cover the most commonly used fairness criterion and can be further generalized and customized for fairness evaluation.

Pengyuan Wang, HyunJun Jung, Yitong Li et al.

Object pose estimation is crucial for robotic applications and augmented reality. Beyond instance level 6D object pose estimation methods, estimating category-level pose and shape has become a promising trend. As such, a new research field needs to be supported by well-designed datasets. To provide a benchmark with high-quality ground truth annotations to the community, we introduce a multimodal dataset for category-level object pose estimation with photometrically challenging objects termed PhoCaL. PhoCaL comprises 60 high quality 3D models of household objects over 8 categories including highly reflective, transparent and symmetric objects. We developed a novel robot-supported multi-modal (RGB, depth, polarisation) data acquisition and annotation process. It ensures sub-millimeter accuracy of the pose for opaque textured, shiny and transparent objects, no motion blur and perfect camera synchronisation. To set a benchmark for our dataset, state-of-the-art RGB-D and monocular RGB methods are evaluated on the challenging scenes of PhoCaL.

Yifan Zhang, Yunjie Gu, Yue Zhu et al.

Grid-following inverters have been widely adopted as a grid interface for renewable energy, and ensuring their small-signal and large-signal stability is critical to modern power systems. Their large-signal, or transient, stability is a significant challenge to analyze because of the interaction of the phase-locked loop (PLL), which must maintain synchronism with various outer-loop controllers. Simple analysis in which outer-loop controllers are idealized is insufficient, and the interactions between the nonlinear dynamics of the PLL and the dynamics of the DC-link voltage control (DVC), as well as the AC terminal voltage control (TVC) when present, must be considered. An asymptotic analysis approach, termed the bandwidth separation method, is proposed. This method enables simplification and order reduction of the original differential equations when sufficient bandwidth separation exists. Through this method, the interaction between the DVC and PLL is explicitly characterized, revealing that such interaction degrades system stability and shrinks the stability region. The analysis also indicates that voltage instability, rather than PLL loss of synchronization alone, is often the root cause of transient instability. Optimal bandwidth configurations for the PLL and DVC are identified under various grid fault conditions: a larger PLL bandwidth improves resilience to phase-jump faults, while a larger DVC bandwidth enhances tolerance to power fluctuations. In addition, the influence of the TVC loop is analyzed, showing that a high TVC bandwidth can mitigate the destabilizing effects of PLL-DVC interaction and further improve transient stability. All analytical findings are validated through hardware-in-the-loop (HIL) experiments.

Matan Atad, Vitalii Dmytrenko, Yitong Li et al.

Deep learning models used in medical image analysis are prone to raising reliability concerns due to their black-box nature. To shed light on these black-box models, previous works predominantly focus on identifying the contribution of input features to the diagnosis, i.e., feature attribution. In this work, we explore counterfactual explanations to identify what patterns the models rely on for diagnosis. Specifically, we investigate the effect of changing features within chest X-rays on the classifier's output to understand its decision mechanism. We leverage a StyleGAN-based approach (StyleEx) to create counterfactual explanations for chest X-rays by manipulating specific latent directions in their latent space. In addition, we propose EigenFind to significantly reduce the computation time of generated explanations. We clinically evaluate the relevancy of our counterfactual explanations with the help of radiologists. Our code is publicly available.

Xin Wang, Yasheng Wang, Yao Wan et al.

Automatically generating compilable programs with (or without) natural language descriptions has always been a touchstone problem for computational linguistics and automated software engineering. Existing deep-learning approaches model code generation as text generation, either constrained by grammar structures in decoder, or driven by pre-trained language models on large-scale code corpus (e.g., CodeGPT, PLBART, and CodeT5). However, few of them account for compilability of the generated programs. To improve compilability of the generated programs, this paper proposes COMPCODER, a three-stage pipeline utilizing compiler feedback for compilable code generation, including language model fine-tuning, compilability reinforcement, and compilability discrimination. Comprehensive experiments on two code generation tasks demonstrate the effectiveness of our proposed approach, improving the success rate of compilation from 44.18 to 89.18 in code completion on average and from 70.3 to 96.2 in text-to-code generation, respectively, when comparing with the state-of-the-art CodeGPT.

HyunJun Jung, Patrick Ruhkamp, Guangyao Zhai et al.

Depth estimation is a core task in 3D computer vision. Recent methods investigate the task of monocular depth trained with various depth sensor modalities. Every sensor has its advantages and drawbacks caused by the nature of estimates. In the literature, mostly mean average error of the depth is investigated and sensor capabilities are typically not discussed. Especially indoor environments, however, pose challenges for some devices. Textureless regions pose challenges for structure from motion, reflective materials are problematic for active sensing, and distances for translucent material are intricate to measure with existing sensors. This paper proposes HAMMER, a dataset comprising depth estimates from multiple commonly used sensors for indoor depth estimation, namely ToF, stereo, structured light together with monocular RGB+P data. We construct highly reliable ground truth depth maps with the help of 3D scanners and aligned renderings. A popular depth estimators is trained on this data and typical depth senosors. The estimates are extensively analyze on different scene structures. We notice generalization issues arising from various sensor technologies in household environments with challenging but everyday scene content. HAMMER, which we make publicly available, provides a reliable base to pave the way to targeted depth improvements and sensor fusion approaches.

Haoqin Tu, Yitong Li

Recent advances in neural-based generative modeling have reignited the hopes of having computer systems capable of conversing with humans and able to understand natural language. The employment of deep neural architectures has been largely explored in a multitude of context and tasks to fulfill various user needs. On one hand, producing textual content that meets specific requirements is of priority for a model to seamlessly conduct conversations with different groups of people. On the other hand, latent variable models (LVM) such as variational auto-encoders (VAEs) as one of the most popular genres of generative models are designed to characterize the distributional pattern of textual data. Thus they are inherently capable of learning the integral textual features that are worth exploring for controllable pursuits. \noindent This overview gives an introduction to existing generation schemes, problems associated with text variational auto-encoders, and a review of several applications about the controllable generation that are instantiations of these general formulations,\footnote{A detailed paper list is available at \url{https://github.com/ImKeTT/CTG-latentAEs}} as well as related datasets, metrics and discussions for future researches. Hopefully, this overview will provide an overview of living questions, popular methodologies and raw thoughts for controllable language generation under the scope of variational auto-encoder.

Martin Rath, Morteza Ghahremani, Yitong Li et al.

Computed tomography (CT) provides rich 3D anatomical details but is often constrained by high radiation exposure, substantial costs, and limited availability. While standard chest X-rays are cost-effective and widely accessible, they only provide 2D projections with limited pathological information. Reconstructing 3D CT volumes from 2D X-rays offers a transformative solution to increase diagnostic accessibility, yet existing methods predominantly rely on synthetic X-ray projections, limiting clinical generalization. In this work, we propose AXON, a multi-stage diffusion-based framework that reconstructs high-fidelity 3D CT volumes directly from real X-rays. AXON employs a coarse-to-fine strategy, with a Brownian Bridge diffusion model-based initial stage for global structural synthesis, followed by a ControlNet-based refinement stage for local intensity optimization. It also supports bi-planar X-ray input to mitigate depth ambiguities inherent in 2D-to-3D reconstruction. A super-resolution network is integrated to upscale the generated volumes to achieve diagnostic-grade resolution. Evaluations on both public and external datasets demonstrate that AXON significantly outperforms state-of-the-art baselines, achieving a 11.9% improvement in PSNR and a 11.0% increase in SSIM with robust generalizability across disparate clinical distributions. Our code is available at https://github.com/ai-med/AXON.

Yitong Li, Igor Yakushev, Dennis M. Hedderich et al.

Positron emission tomography (PET) is a widely recognized technique for diagnosing neurodegenerative diseases, offering critical functional insights. However, its high costs and radiation exposure hinder its widespread use. In contrast, magnetic resonance imaging (MRI) does not involve such limitations. While MRI also detects neurodegenerative changes, it is less sensitive for diagnosis compared to PET. To overcome such limitations, one approach is to generate synthetic PET from MRI. Recent advances in generative models have paved the way for cross-modality medical image translation; however, existing methods largely emphasize structural preservation while neglecting the critical need for pathology awareness. To address this gap, we propose PASTA, a novel image translation framework built on conditional diffusion models with enhanced pathology awareness. PASTA surpasses state-of-the-art methods by preserving both structural and pathological details through its highly interactive dual-arm architecture and multi-modal condition integration. Additionally, we introduce a novel cycle exchange consistency and volumetric generation strategy that significantly enhances PASTA's ability to produce high-quality 3D PET images. Our qualitative and quantitative results demonstrate the high quality and pathology awareness of the synthesized PET scans. For Alzheimer's diagnosis, the performance of these synthesized scans improves over MRI by 4%, almost reaching the performance of actual PET. Our code is available at https://github.com/ai-med/PASTA.

Bin Sun, Yitong Li, Fei Mi et al.

Conditional variational models, using either continuous or discrete latent variables, are powerful for open-domain dialogue response generation. However, previous works show that continuous latent variables tend to reduce the coherence of generated responses. In this paper, we also found that discrete latent variables have difficulty capturing more diverse expressions. To tackle these problems, we combine the merits of both continuous and discrete latent variables and propose a Hybrid Latent Variable (HLV) method. Specifically, HLV constrains the global semantics of responses through discrete latent variables and enriches responses with continuous latent variables. Thus, we diversify the generated responses while maintaining relevance and coherence. In addition, we propose Conditional Hybrid Variational Transformer (CHVT) to construct and to utilize HLV with transformers for dialogue generation. Through fine-grained symbolic-level semantic information and additive Gaussian mixing, we construct the distribution of continuous variables, prompting the generation of diverse expressions. Meanwhile, to maintain the relevance and coherence, the discrete latent variable is optimized by self-separation training. Experimental results on two dialogue generation datasets (DailyDialog and Opensubtitles) show that CHVT is superior to traditional transformer-based variational mechanism w.r.t. diversity, relevance and coherence metrics. Moreover, we also demonstrate the benefit of applying HLV to fine-tuning two pre-trained dialogue models (PLATO and BART-base).

Kai Chen, Ye Wang, Yitong Li et al.

Temporal factors are tied to the growth of facts in realistic applications, such as the progress of diseases and the development of political situation, therefore, research on Temporal Knowledge Graph (TKG) attracks much attention. In TKG, relation patterns inherent with temporality are required to be studied for representation learning and reasoning across temporal facts. However, existing methods can hardly model temporal relation patterns, nor can capture the intrinsic connections between relations when evolving over time, lacking of interpretability. In this paper, we propose a novel temporal modeling method which represents temporal entities as Rotations in Quaternion Vector Space (RotateQVS) and relations as complex vectors in Hamilton's quaternion space. We demonstrate our method can model key patterns of relations in TKG, such as symmetry, asymmetry, inverse, and can further capture time-evolved relations by theory. Empirically, we show that our method can boost the performance of link prediction tasks over four temporal knowledge graph benchmarks.

Yifan Zhang, Yaoxin Wang, Yunjie Gu et al.

Grid-following (GFL) inverters have very different large-signal stability characteristics to synchronous generators, and convenient concepts such as the equal-area criterion and global energy function do not apply in the same way. Existing studies mainly focus on the synchronization stability of an individual GFL inverter, while interactions between multiple inverters are less often addressed. This paper elucidates the interaction mechanisms between heterogeneous inverters, covering GFL, grid-forming (GFM), and grid-supporting (GSP) types, to determine the stability boundaries of systems with mixed inverter compositions. The generalized large-signal model for two-inverter systems is derived for various inverter combinations. This paper establishes that systems containing GFL inverters do not admit a global energy function, fundamentally limiting the applicability of traditional direct methods. To overcome this barrier, a manifold method is employed to accurately determine the region of attraction (ROA). To address the computational complexity of the manifold method, reduced-order models of inverter are used based on multiscale analysis. The large-signal stability margin is assessed by the shortest distance from a stable equilibrium point (SEP) to the boundary of the ROA, which is called the stability radius (SR). Using the proposed framework, the analysis reults of two-inverter system show that both GFM and GSP inverters significantly enhance the large-signal stability of a two-inverter system where the other inverter is GFL, with GFM providing slightly superior performance. This improvement is attributed to the voltage support effects and is maximized when the GFM or GSP inverter is located at the midpoint of the transmission line, where the voltage is lowest. All findings in this paper are validated through both EMT simulations and power hardware-in-the-loop (PHIL) experiments.

Bin Sun, Shaoxiong Feng, Yiwei Li et al.

Complex dialogue mappings (CDM), including one-to-many and many-to-one mappings, tend to make dialogue models generate incoherent or dull responses, and modeling these mappings remains a huge challenge for neural dialogue systems. To alleviate these problems, methods like introducing external information, reconstructing the optimization function, and manipulating data samples are proposed, while they primarily focus on avoiding training with CDM, inevitably weakening the model's ability of understanding CDM in human conversations and limiting further improvements in model performance. This paper proposes a Sentence Semantic \textbf{Seg}mentation guided \textbf{C}onditional \textbf{V}ariational \textbf{A}uto-\textbf{E}ncoder (SegCVAE) method which can model and take advantages of the CDM data. Specifically, to tackle the incoherent problem caused by one-to-many, SegCVAE uses response-related prominent semantics to constrained the latent variable. To mitigate the non-diverse problem brought by many-to-one, SegCVAE segments multiple prominent semantics to enrich the latent variables. Three novel components, Internal Separation, External Guidance, and Semantic Norms, are proposed to achieve SegCVAE. On dialogue generation tasks, both the automatic and human evaluation results show that SegCVAE achieves new state-of-the-art performance.

Yang Xu, Yunlong Feng, Honglin Mu et al.

Through reading the documentation in the context, tool-using language models can dynamically extend their capability using external tools. The cost is that we have to input lengthy documentation every time the model needs to use the tool, occupying the input window as well as slowing down the decoding process. Given the progress in general-purpose compression, soft context compression is a suitable approach to alleviate the problem. However, when compressing tool documentation, existing methods suffer from the weaknesses of key information loss (specifically, tool/parameter name errors) and difficulty in adjusting the length of compressed sequences based on documentation lengths. To address these problems, we propose two strategies for compressing tool documentation into concise and precise summary sequences for tool-using language models. 1) Selective compression strategy mitigates key information loss by deliberately retaining key information as raw text tokens. 2) Block compression strategy involves dividing tool documentation into short chunks and then employing a fixed-length compression model to achieve variable-length compression. This strategy facilitates the flexible adjustment of the compression ratio. Results on API-Bank and APIBench show that our approach reaches a performance comparable to the upper-bound baseline under up to 16x compression ratio.

Yitong Li, Junsong Chen, Shuchen Xue et al.

Reinforcement-Learning-based post-training has recently emerged as a promising paradigm for aligning text-to-image diffusion models with human preferences. In recent studies, increasing the rollout group size yields pronounced performance improvements, indicating substantial room for further alignment gains. However, scaling rollouts on large-scale foundational diffusion models (e.g., FLUX.1-12B) imposes a heavy computational burden. To alleviate this bottleneck, we explore the integration of FP4 quantization into Diffusion RL rollouts. Yet, we identify that naive quantized pipelines inherently introduce risks of performance degradation. To overcome this dilemma between efficiency and training integrity, we propose Sol-RL (Speed-of-light RL), a novel FP4-empowered Two-stage Reinforcement Learning framework. First, we utilize high-throughput NVFP4 rollouts to generate a massive candidate pool and extract a highly contrastive subset. Second, we regenerate these selected samples in BF16 precision and optimize the policy exclusively on them. By decoupling candidate exploration from policy optimization, Sol-RL integrates the algorithmic mechanisms of rollout scaling with the system-level throughput gains of NVFP4. This synergistic algorithm-hardware design effectively accelerates the rollout phase while reserving high-fidelity samples for optimization. We empirically demonstrate that our framework maintains the training integrity of BF16 precision pipeline while fully exploiting the throughput gains enabled by FP4 arithmetic. Extensive experiments across SANA, FLUX.1, and SD3.5-L substantiate that our approach delivers superior alignment performance across multiple metrics while accelerating training convergence by up to $4.64\times$, unlocking the power of massive rollout scaling at a fraction of the cost.

Haoran Lu, Ruihai Wu, Yitong Li et al.

Manipulating garments and fabrics has long been a critical endeavor in the development of home-assistant robots. However, due to complex dynamics and topological structures, garment manipulations pose significant challenges. Recent successes in reinforcement learning and vision-based methods offer promising avenues for learning garment manipulation. Nevertheless, these approaches are severely constrained by current benchmarks, which offer limited diversity of tasks and unrealistic simulation behavior. Therefore, we present GarmentLab, a content-rich benchmark and realistic simulation designed for deformable object and garment manipulation. Our benchmark encompasses a diverse range of garment types, robotic systems and manipulators. The abundant tasks in the benchmark further explores of the interactions between garments, deformable objects, rigid bodies, fluids, and human body. Moreover, by incorporating multiple simulation methods such as FEM and PBD, along with our proposed sim-to-real algorithms and real-world benchmark, we aim to significantly narrow the sim-to-real gap. We evaluate state-of-the-art vision methods, reinforcement learning, and imitation learning approaches on these tasks, highlighting the challenges faced by current algorithms, notably their limited generalization capabilities. Our proposed open-source environments and comprehensive analysis show promising boost to future research in garment manipulation by unlocking the full potential of these methods. We guarantee that we will open-source our code as soon as possible. You can watch the videos in supplementary files to learn more about the details of our work. Our project page is available at: https://garmentlab.github.io/

Yitong Li, Morteza Ghahremani, Youssef Wally et al.

Diagnosing dementia, particularly for Alzheimer's Disease (AD) and frontotemporal dementia (FTD), is complex due to overlapping symptoms. While magnetic resonance imaging (MRI) and positron emission tomography (PET) data are critical for the diagnosis, integrating these modalities in deep learning faces challenges, often resulting in suboptimal performance compared to using single modalities. Moreover, the potential of multi-modal approaches in differential diagnosis, which holds significant clinical importance, remains largely unexplored. We propose a novel framework, DiaMond, to address these issues with vision Transformers to effectively integrate MRI and PET. DiaMond is equipped with self-attention and a novel bi-attention mechanism that synergistically combine MRI and PET, alongside a multi-modal normalization to reduce redundant dependency, thereby boosting the performance. DiaMond significantly outperforms existing multi-modal methods across various datasets, achieving a balanced accuracy of 92.4% in AD diagnosis, 65.2% for AD-MCI-CN classification, and 76.5% in differential diagnosis of AD and FTD. We also validated the robustness of DiaMond in a comprehensive ablation study. The code is available at https://github.com/ai-med/DiaMond.

Yitong Li, Tom Nuno Wolf, Sebastian Pölsterl et al.

Differential diagnosis of dementia is challenging due to overlapping symptoms, with structural magnetic resonance imaging (MRI) being the primary method for diagnosis. Despite the clinical value of computer-aided differential diagnosis, research has been limited, mainly due to the absence of public datasets that contain diverse types of dementia. This leaves researchers with small in-house datasets that are insufficient for training deep neural networks (DNNs). Self-supervised learning shows promise for utilizing unlabeled MRI scans in training, but small batch sizes for volumetric brain scans make its application challenging. To address these issues, we propose Triplet Training for differential diagnosis with limited target data. It consists of three key stages: (i) self-supervised pre-training on unlabeled data with Barlow Twins, (ii) self-distillation on task-related data, and (iii) fine-tuning on the target dataset. Our approach significantly outperforms traditional training strategies, achieving a balanced accuracy of 75.6%. We further provide insights into the training process by visualizing changes in the latent space after each step. Finally, we validate the robustness of Triplet Training in terms of its individual components in a comprehensive ablation study. Our code is available at https://github.com/ai-med/TripletTraining.

Fabian Bongratz, Yitong Li, Sama Elbaroudy et al.

Synthetic neuroimaging data can mitigate critical limitations of real-world datasets, including the scarcity of rare phenotypes, domain shifts across scanners, and insufficient longitudinal coverage. However, existing generative models largely rely on weak conditioning signals, such as labels or text, which lack anatomical grounding and often produce biologically implausible outputs. To this end, we introduce Cor2Vox, a cortex-grounded generative framework for brain magnetic resonance image (MRI) synthesis that ties image generation to continuous structural priors of the cerebral cortex. It leverages high-resolution cortical surfaces to guide a 3D shape-to-image Brownian bridge diffusion process, enabling topologically faithful synthesis and precise control over underlying anatomies. To support the generation of new, realistic brain shapes, we developed a large-scale statistical shape model of cortical morphology derived from over 33,000 UK Biobank scans. We validated the fidelity of Cor2Vox based on traditional image quality metrics, advanced cortical surface reconstruction, and whole-brain segmentation quality, outperforming many baseline methods. Across three applications, namely (i) anatomically consistent synthesis, (ii) simulation of progressive gray matter atrophy, and (iii) harmonization of in-house frontotemporal dementia scans with public datasets, Cor2Vox preserved fine-grained cortical morphology at the sub-voxel level, exhibiting remarkable robustness to variations in cortical geometry and disease phenotype without retraining.

Dianyi Wang, Siyuan Wang, Zejun Li et al.

Large Vision-Language Models (LVLMs) have demonstrated remarkable performance across multi-modal tasks by scaling model size and training data. However, these dense LVLMs incur significant computational costs and motivate the exploration of sparse Mixture of Experts (MoE) architectures. While MoE improve parameter efficiency, effectively applying MoE to simultaneously model modality-specific features and cross-modal associations in LVLMs remains challenging. In this work, we propose to incorporate Mixture of Intra- and Inter-Modality Experts (MoIIE) to LVLMs. For each token, expert routing is guided by its modality, directing tokens to their respective intra-modality experts as well as a shared pool of inter-modality experts, enabling the model to jointly learn rich intra-modal features and cross-modal interactions. We further introduce an effective and straightforward two-stage training strategy, which facilitates the direct activation of both MoE and multi-modal capabilities. Extensive experiments across different data scales and LLM backbone demonstrate the effectiveness, efficiency and generality of our approach. Notably, our MoIIE models with 5.5B and 11.3B activated parameters match or even surpass the performance of existing advanced open-source MoE-LLMs based multi-modal models that involve more activated parameters. The code is available at https://github.com/AlenjandroWang/MoIIE.

Fabian Bongratz, Yitong Li, Sama Elbaroudy et al.

Despite recent advances in medical image generation, existing methods struggle to produce anatomically plausible 3D structures. In synthetic brain magnetic resonance images (MRIs), characteristic fissures are often missing, and reconstructed cortical surfaces appear scattered rather than densely convoluted. To address this issue, we introduce Cor2Vox, the first diffusion model-based method that translates continuous cortical shape priors to synthetic brain MRIs. To achieve this, we leverage a Brownian bridge process which allows for direct structured mapping between shape contours and medical images. Specifically, we adapt the concept of the Brownian bridge diffusion model to 3D and extend it to embrace various complementary shape representations. Our experiments demonstrate significant improvements in the geometric accuracy of reconstructed structures compared to previous voxel-based approaches. Moreover, Cor2Vox excels in image quality and diversity, yielding high variation in non-target structures like the skull. Finally, we highlight the capability of our approach to simulate cortical atrophy at the sub-voxel level. Our code is available at https://github.com/ai-med/Cor2Vox.

Bing Wu, Chang Zou, Changlin Li et al.

We present HunyuanVideo 1.5, a lightweight yet powerful open-source video generation model that achieves state-of-the-art visual quality and motion coherence with only 8.3 billion parameters, enabling efficient inference on consumer-grade GPUs. This achievement is built upon several key components, including meticulous data curation, an advanced DiT architecture featuring selective and sliding tile attention (SSTA), enhanced bilingual understanding through glyph-aware text encoding, progressive pre-training and post-training, and an efficient video super-resolution network. Leveraging these designs, we developed a unified framework capable of high-quality text-to-video and image-to-video generation across multiple durations and resolutions. Extensive experiments demonstrate that this compact and proficient model establishes a new state-of-the-art among open-source video generation models. By releasing the code and model weights, we provide the community with a high-performance foundation that lowers the barrier to video creation and research, making advanced video generation accessible to a broader audience. All open-source assets are publicly available at https://github.com/Tencent-Hunyuan/HunyuanVideo-1.5.

Yitong Li, Ralph Buchert, Benita Schmitz-Koep et al.

Positron emission tomography (PET) with 18F-Fluorodeoxyglucose (FDG) is an established tool in the diagnostic workup of patients with suspected dementing disorders. However, compared to the routinely available magnetic resonance imaging (MRI), FDG-PET remains significantly less accessible and substantially more expensive. Here, we present SiM2P, a 3D diffusion bridge-based framework that learns a probabilistic mapping from MRI and auxiliary patient information to simulate FDG-PET images of diagnostic quality. In a blinded clinical reader study, two neuroradiologists and two nuclear medicine physicians rated the original MRI and SiM2P-simulated PET images of patients with Alzheimer's disease, behavioral-variant frontotemporal dementia, and cognitively healthy controls. SiM2P significantly improved the overall diagnostic accuracy of differentiating between three groups from 75.0% to 84.7% (p<0.05). Notably, the simulated PET images received higher diagnostic certainty ratings and achieved superior interrater agreement compared to the MRI images. Finally, we developed a practical workflow for local deployment of the SiM2P framework. It requires as few as 20 site-specific cases and only basic demographic information. This approach makes the established diagnostic benefits of FDG-PET imaging more accessible to patients with suspected dementing disorders, potentially improving early detection and differential diagnosis in resource-limited settings. Our code is available at https://github.com/Yiiitong/SiM2P.

Jingcong Liang, Siyuan Wang, Miren Tian et al.

Mixture-of-Experts (MoE) enables efficient scaling of large language models (LLMs) with sparsely activated experts during inference. To effectively deploy large MoE models on memory-constrained devices, many systems introduce *expert offloading* that caches a subset of experts in fast memory, leaving others on slow memory to run on CPU or load on demand. While some research has exploited the locality of expert activations, where consecutive tokens activate similar experts, the degree of this **local routing consistency** varies across models and remains understudied. In this paper, we propose two metrics to measure local routing consistency of MoE models: (1) **Segment Routing Best Performance (SRP)**, which evaluates how well a fixed group of experts can cover the needs of a segment of tokens, and (2) **Segment Cache Best Hit Rate (SCH)**, which measures the optimal segment-level cache hit rate under a given cache size limit. We analyzed 20 MoE LLMs with diverse sizes and architectures and found that models that apply MoE on every layer and do not use shared experts exhibit the highest local routing consistency. We further showed that domain-specialized experts contribute more to routing consistency than vocabulary-specialized ones, and that most models can balance between cache effectiveness and efficiency with cache sizes approximately 2x the active experts. These findings pave the way for memory-efficient MoE design and deployment without compromising inference speed. We publish the code for replicating experiments at https://github.com/ljcleo/moe-lrc .

Yitong Li, Morteza Ghahremani, Christian Wachinger

Recent vision-language foundation models deliver state-of-the-art results on natural image classification but falter on medical images due to pronounced domain shifts. At the same time, training a medical foundation model requires substantial resources, including extensive annotated data and high computational capacity. To bridge this gap with minimal overhead, we introduce MedBridge, a lightweight multimodal adaptation framework that re-purposes pretrained VLMs for accurate medical image diagnosis. MedBridge comprises three key components. First, a Focal Sampling module that extracts high-resolution local regions to capture subtle pathological features and compensate for the limited input resolution of general-purpose VLMs. Second, a Query Encoder (QEncoder) injects a small set of learnable queries that attend to the frozen feature maps of VLM, aligning them with medical semantics without retraining the entire backbone. Third, a Mixture of Experts mechanism, driven by learnable queries, harnesses the complementary strength of diverse VLMs to maximize diagnostic performance. We evaluate MedBridge on five medical imaging benchmarks across three key adaptation tasks, demonstrating its superior performance in both cross-domain and in-domain adaptation settings, even under varying levels of training data availability. Notably, MedBridge achieved over 6-15% improvement in AUC compared to state-of-the-art VLM adaptation methods in multi-label thoracic disease diagnosis, underscoring its effectiveness in leveraging foundation models for accurate and data-efficient medical diagnosis. Our code is available at https://github.com/ai-med/MedBridge.

Xiaofei Yu, Yitong Li, Jie Ma

Remote sensing image change captioning (RSICC) aims at generating human-like language to describe the semantic changes between bi-temporal remote sensing image pairs. It provides valuable insights into environmental dynamics and land management. Unlike conventional change captioning task, RSICC involves not only retrieving relevant information across different modalities and generating fluent captions, but also mitigating the impact of pixel-level differences on terrain change localization. The pixel problem due to long time span decreases the accuracy of generated caption. Inspired by the remarkable generative power of diffusion model, we propose a probabilistic diffusion model for RSICC to solve the aforementioned problems. In training process, we construct a noise predictor conditioned on cross modal features to learn the distribution from the real caption distribution to the standard Gaussian distribution under the Markov chain. Meanwhile, a cross-mode fusion and a stacking self-attention module are designed for noise predictor in the reverse process. In testing phase, the well-trained noise predictor helps to estimate the mean value of the distribution and generate change captions step by step. Extensive experiments on the LEVIR-CC dataset demonstrate the effectiveness of our Diffusion-RSCC and its individual components. The quantitative results showcase superior performance over existing methods across both traditional and newly augmented metrics. The code and materials will be available online at https://github.com/Fay-Y/Diffusion-RSCC.

Jiajun Wang, Morteza Ghahremani, Yitong Li et al.

Controllable text-to-image (T2I) diffusion models have shown impressive performance in generating high-quality visual content through the incorporation of various conditions. Current methods, however, exhibit limited performance when guided by skeleton human poses, especially in complex pose conditions such as side or rear perspectives of human figures. To address this issue, we present Stable-Pose, a novel adapter model that introduces a coarse-to-fine attention masking strategy into a vision Transformer (ViT) to gain accurate pose guidance for T2I models. Stable-Pose is designed to adeptly handle pose conditions within pre-trained Stable Diffusion, providing a refined and efficient way of aligning pose representation during image synthesis. We leverage the query-key self-attention mechanism of ViTs to explore the interconnections among different anatomical parts in human pose skeletons. Masked pose images are used to smoothly refine the attention maps based on target pose-related features in a hierarchical manner, transitioning from coarse to fine levels. Additionally, our loss function is formulated to allocate increased emphasis to the pose region, thereby augmenting the model's precision in capturing intricate pose details. We assessed the performance of Stable-Pose across five public datasets under a wide range of indoor and outdoor human pose scenarios. Stable-Pose achieved an AP score of 57.1 in the LAION-Human dataset, marking around 13% improvement over the established technique ControlNet. The project link and code is available at https://github.com/ai-med/StablePose.

Haoqin Tu, Yitong Li, Fei Mi et al.

Incorporating visual knowledge into text-only dialogue systems has become a potential direction to imitate the way humans think, imagine, and communicate. However, existing multimodal dialogue systems are either confined by the scale and quality of available datasets or the coarse concept of visual knowledge. To address these issues, we provide a new paradigm of constructing multimodal dialogues as well as two datasets extended from text-only dialogues under such paradigm (ReSee-WoW, ReSee-DD). We propose to explicitly split the visual knowledge into finer granularity (``turn-level'' and ``entity-level''). To further boost the accuracy and diversity of augmented visual information, we retrieve them from the Internet or a large image dataset. To demonstrate the superiority and universality of the provided visual knowledge, we propose a simple but effective framework ReSee to add visual representation into vanilla dialogue models by modality concatenations. We also conduct extensive experiments and ablations w.r.t. different model configurations and visual knowledge settings. Empirical, encouraging results not only demonstrate the effectiveness of introducing visual knowledge at both entity and turn level but also verify the proposed model ReSee outperforms several state-of-the-art methods on automatic and human evaluations. By leveraging text and vision knowledge, ReSee can produce informative responses with real-world visual concepts. Our code is available at https://github.com/ImKeTT/ReSee.

Yitong Li, Chang Liu, Jie Ma

Weakly supervised learning based on scribble annotations in target extraction of remote sensing images has drawn much interest due to scribbles' flexibility in denoting winding objects and low cost of manually labeling. However, scribbles are too sparse to identify object structure and detailed information, bringing great challenges in target localization and boundary description. To alleviate these problems, in this paper, we construct two inner structure-constraints, a deformation consistency loss and a trainable active contour loss, together with a scribble-constraint to supervise the optimization of the encoder-decoder network without introducing any auxiliary module or extra operation based on prior cues. Comprehensive experiments demonstrate our method's superiority over five state-of-the-art algorithms in this field. Source code is available at https://github.com/yitongli123/ISC-TE.

Fei Mi, Yitong Li, Yulong Zeng et al.

In this paper, we introduce PanGu-Bot, a Chinese pre-trained open-domain dialogue generation model based on a large pre-trained language model (PLM) PANGU-alpha (Zeng et al.,2021). Different from other pre-trained dialogue models trained over a massive amount of dialogue data from scratch, we aim to build a powerful dialogue model with relatively fewer data and computation costs by inheriting valuable language capabilities and knowledge from PLMs. To this end, we train PanGu-Bot from the large PLM PANGU-alpha, which has been proven well-performed on a variety of Chinese natural language tasks. We investigate different aspects of responses generated by PanGu-Bot, including response quality, knowledge, and safety. We show that PanGu-Bot outperforms state-of-the-art Chinese dialogue systems (CDIALGPT (Wang et al., 2020), EVA (Zhou et al., 2021), EVA2.0 (Gu et al., 2022)) w.r.t. the above three aspects. We also demonstrate that PanGu-Bot can be easily deployed to generate emotional responses without further training. Throughout our empirical analysis, we also point out that the PanGu-Bot response quality, knowledge correctness, and safety are still far from perfect, and further explorations are indispensable to building reliable and smart dialogue systems. Our model and code will be available at https://github.com/huawei-noah/Pretrained-Language-Model/tree/master/PanGu-Bot soon.

Lang Cao, Hui Ruan, Yongqian Li et al.

Reinforcement learning with group-based objectives, such as Group Relative Policy Optimization (GRPO), is a common framework for aligning large language models on complex reasoning tasks. However, standard GRPO treats each rollout trajectory as an independent flat sequence and assigns a single sequence-level advantage to all tokens, which leads to sample inefficiency and a length bias toward verbose, redundant chains of thought without improving logical depth. We introduce TreeAdv (Tree-Structured Advantage Redistribution for Group-Based RL), which makes the tree structure of group rollouts explicit for both exploration and advantage assignment. Specifically, TreeAdv builds a group of trees (a forest) based on an entropy-driven sampling method where each tree branches at high-uncertainty decisions while sharing low-uncertainty tokens across rollouts. Then, TreeAdv aggregates token-level advantages for internal tree segments by redistributing the advantages of complete rollouts (all leaf nodes), and TreeAdv can easily apply to group-based objectives such as GRPO or GSPO. Across 10 math reasoning benchmarks, TreeAdv consistently outperforms GRPO and GSPO, while using substantially fewer generated tokens under identical supervision, data, and decoding budgets.

Yifan Zhang, Hsiao-Dong Chiang, Yitong Li et al.

The increasing penetration of inverter-based resources (IBRs) has fundamentally altered the transient stability characteristics of modern power systems. IBRs typically rely on proportional--integral (PI) controllers for synchronization and regulation, resulting in nonlinear swing equations that differ significantly from those of synchronous generators (SGs) and exhibit state-dependent damping. Consequently, although the classical energy function is often adopted in IBR analysis by analogy with SGs, it cannot be directly applied to IBRs with PI controller. A new energy function explicitly tailored to PI controller is proposed in this letter. It admits a unified form and can be applied to a class of nonlinear systems with PI controllers. Two representative cases are considered, including a grid-following (GFL) inverter and a DC-voltage-controlled grid-forming (GFM) inverter, demonstrating less conservative and more effective estimation of the region of attraction (ROA). All findings are verified through hardware-in-the-loop (HIL) experiments.

Yuzhao Heng, Chunyuan Deng, Yitong Li et al. · gatech

Although Large Language Models (LLMs) exhibit remarkable adaptability across domains, these models often fall short in structured knowledge extraction tasks such as named entity recognition (NER). This paper explores an innovative, cost-efficient strategy to harness LLMs with modest NER capabilities for producing superior NER datasets. Our approach diverges from the basic class-conditional prompts by instructing LLMs to self-reflect on the specific domain, thereby generating domain-relevant attributes (such as category and emotions for movie reviews), which are utilized for creating attribute-rich training data. Furthermore, we preemptively generate entity terms and then develop NER context data around these entities, effectively bypassing the LLMs' challenges with complex structures. Our experiments across both general and niche domains reveal significant performance enhancements over conventional data generation methods while being more cost-effective than existing alternatives.

Qingyang Wu, Ying Xu, Tingsong Xiao et al.

Large Language Models (LLMs) have emerged as potent tools for advancing the United Nations' Sustainable Development Goals (SDGs). However, the attitudinal disparities between LLMs and humans towards these goals can pose significant challenges. This study conducts a comprehensive review and analysis of the existing literature on the attitudes of LLMs towards the 17 SDGs, emphasizing the comparison between their attitudes and support for each goal and those of humans. We examine the potential disparities, primarily focusing on aspects such as understanding and emotions, cultural and regional differences, task objective variations, and factors considered in the decision-making process. These disparities arise from the underrepresentation and imbalance in LLM training data, historical biases, quality issues, lack of contextual understanding, and skewed ethical values reflected. The study also investigates the risks and harms that may arise from neglecting the attitudes of LLMs towards the SDGs, including the exacerbation of social inequalities, racial discrimination, environmental destruction, and resource wastage. To address these challenges, we propose strategies and recommendations to guide and regulate the application of LLMs, ensuring their alignment with the principles and goals of the SDGs, and therefore creating a more just, inclusive, and sustainable future.

Guang Yin, Yitong Li, Yixuan Wang et al.

Natural language instructions for robotic manipulation tasks often exhibit ambiguity and vagueness. For instance, the instruction "Hang a mug on the mug tree" may involve multiple valid actions if there are several mugs and branches to choose from. Existing language-conditioned policies typically rely on end-to-end models that jointly handle high-level semantic understanding and low-level action generation, which can result in suboptimal performance due to their lack of modularity and interpretability. To address these challenges, we introduce a novel robotic manipulation framework that can accomplish tasks specified by potentially ambiguous natural language. This framework employs a Vision-Language Model (VLM) to interpret abstract concepts in natural language instructions and generates task-specific code - an interpretable and executable intermediate representation. The generated code interfaces with the perception module to produce 3D attention maps that highlight task-relevant regions by integrating spatial and semantic information, effectively resolving ambiguities in instructions. Through extensive experiments, we identify key limitations of current imitation learning methods, such as poor adaptation to language and environmental variations. We show that our approach excels across challenging manipulation tasks involving language ambiguity, contact-rich manipulation, and multi-object interactions.

Honglin Mu, Yang Xu, Yunlong Feng et al.

With the rise of Large Language Models (LLMs), AI assistants' ability to utilize tools, especially through API calls, has advanced notably. This progress has necessitated more accurate evaluation methods. Many existing studies adopt static evaluation, where they assess AI assistants' API call based on pre-defined dialogue histories. However, such evaluation method can be misleading, as an AI assistant might fail in generating API calls from preceding human interaction in real cases. Instead of the resource-intensive method of direct human-machine interactions, we propose Automated Dynamic Evaluation (AutoDE) to assess an assistant's API call capability without human involvement. In our framework, we endeavor to closely mirror genuine human conversation patterns in human-machine interactions, using a LLM-based user agent, equipped with a user script to ensure human alignment. Experimental results highlight that AutoDE uncovers errors overlooked by static evaluations, aligning more closely with human assessment. Testing four AI assistants using our crafted benchmark, our method further mirrored human evaluation compared to conventional static evaluations.

Tiezheng Zhang, Yitong Li, Yu-cheng Chou et al.

Building state-of-the-art Vision-Language Models (VLMs) with strong captioning capabilities typically necessitates training on billions of high-quality image-text pairs, requiring millions of GPU hours. This paper introduces the Vision-Language-Vision (VLV) auto-encoder framework, which strategically leverages key pretrained components: a vision encoder, the decoder of a Text-to-Image (T2I) diffusion model, and subsequently, a Large Language Model (LLM). Specifically, we establish an information bottleneck by regularizing the language representation space, achieved through freezing the pretrained T2I diffusion decoder. Our VLV pipeline effectively distills knowledge from the text-conditioned diffusion model using continuous embeddings, demonstrating comprehensive semantic understanding via high-quality reconstructions. Furthermore, by fine-tuning a pretrained LLM to decode the intermediate language representations into detailed descriptions, we construct a state-of-the-art (SoTA) captioner comparable to leading models like GPT-4o and Gemini 2.0 Flash. Our method demonstrates exceptional cost-efficiency and significantly reduces data requirements; by primarily utilizing single-modal images for training and maximizing the utility of existing pretrained models (image encoder, T2I diffusion model, and LLM), it circumvents the need for massive paired image-text datasets, keeping the total training expenditure under $1,000 USD.

Shibo Jie, Yehui Tang, Kai Han et al. · pku

Mixture-of-Experts (MoE) activates only a subset of experts during inference, allowing the model to maintain low inference FLOPs and latency even as the parameter count scales up. However, since MoE dynamically selects the experts, all the experts need to be loaded into VRAM. Their large parameter size still limits deployment, and offloading, which load experts into VRAM only when needed, significantly increase inference latency. To address this, we propose Mixture of Lookup Experts (MoLE), a new MoE architecture that is efficient in both communication and VRAM usage. In MoLE, the experts are Feed-Forward Networks (FFNs) during training, taking the output of the embedding layer as input. Before inference, these experts can be re-parameterized as lookup tables (LUTs) that retrieves expert outputs based on input ids, and offloaded to storage devices. Therefore, we do not need to perform expert computations during inference. Instead, we directly retrieve the expert's computation results based on input ids and load them into VRAM, and thus the resulting communication overhead is negligible. Experiments show that, with the same FLOPs and VRAM usage, MoLE achieves inference speeds comparable to dense models and significantly faster than MoE with experts offloading, while maintaining performance on par with MoE.

Zhenshuai Liu, Yitong Li, Zirui Wang et al.

There has been widespread global increasing use of renewable energy sources, which are usually connected to the electricity grids via power electronic inverters. Traditionally, these inverter-based resources operate in either grid-forming (GFM) or grid-following (GFL) mode. But more recently, the need of switching between these two modes are glowingly required because of the complex operation scenarios of systems such as source-side limitations, grid-side services, fault disturbances, etc. However, due to the differences between GFM and GFL modes, a direct switching between them would lead to large oscillations or even instability of inverters. Therefore, in this paper, a method called state mapping method for analyzing the switching transient and designing the switching control is proposed. Based on this method, an ideally-smooth transition between GFM and GFL can be achieved. The effectiveness of the proposed method is verified by both the theoretical analysis and experiment tests.

Zirui Wang, Yitong Li, Quanchi Wu et al.

Grid-connected inverters are required to operate stably under a wide range of grid conditions. However, conventional grid-following (GFL) control may suffer from instability under weak-grid conditions, while grid-forming (GFM) control may exhibit unstable oscillations under strong-grid conditions. To address these issues, a hybrid voltage-current control method is proposed in this article. A voltage control is introduced on the d-axis, while a current control is adopted on the q-axis, enabling the inverter to exhibit voltage-source characteristics on the d-axis and current-source characteristics on the q-axis. In this way, the proposed control integrates the characteristics of both conventional GFL and GFM control. A full-order model is established to analyze the port characteristics and small-signal stability of the systems. Finally, the effectiveness of the proposed control strategy is validated through simulations and experiments on a 1.5 kW inverter experimental platform. The results show that the proposed control maintains stable operation under different grid conditions with varying short-circuit ratios (SCRs).

Haotian Sun, Yitong Li, Yuchen Zhuang et al.

Contrastive Language-Image Pretraining (CLIP) has demonstrated strong zero-shot performance across diverse downstream text-image tasks. Existing CLIP methods typically optimize a contrastive objective using negative samples drawn from each minibatch. To achieve robust representation learning, these methods require extremely large batch sizes and escalate computational demands to hundreds or even thousands of GPUs. Prior approaches to mitigate this issue often compromise downstream performance, prolong training duration, or face scalability challenges with very large datasets. To overcome these limitations, we propose AmorLIP, an efficient CLIP pretraining framework that amortizes expensive computations involved in contrastive learning through lightweight neural networks, which substantially improves training efficiency and performance. Leveraging insights from a spectral factorization of energy-based models, we introduce novel amortization objectives along with practical techniques to improve training stability. Extensive experiments across 38 downstream tasks demonstrate the superior zero-shot classification and retrieval capabilities of AmorLIP, consistently outperforming standard CLIP baselines with substantial relative improvements of up to 12.24%.

Lang Cao, Renhong Chen, Yingtian Zou et al.

We introduce the Entropy-Driven Uncertainty Process Reward Model (EDU-PRM), a novel entropy-driven training framework for process reward modeling that enables dynamic, uncertainty-aligned segmentation of complex reasoning steps, eliminating the need for costly manual step annotations. Unlike previous Process Reward Models (PRMs) that rely on static partitioning and human labeling, EDU-PRM automatically anchors step boundaries at tokens with high predictive entropy, effectively capturing intrinsic logical transitions and facilitating efficient exploration of diverse reasoning paths. On the ProcessBench benchmark, EDU-PRM outperforms strong public PRM baselines, such as Math-Shepherd PRM and Omega PRM, and EDU-PRM achieves comparable results with SOTA models while only using 1.5% training data. Furthermore, by leveraging our proposed EDU sampling strategy, we observe accuracy boosts from 64.7% to 67.3% for generative reasoning tasks, accompanied by a reduction of 32% in token usage. These findings underscore the potential of EDU-PRM as a scalable and annotation-efficient paradigm for process supervision in mathematical reasoning, paving the way for more efficient and robust approaches to complex mathematical problem solving.

Bailiang Jian, Jiazhen Pan, Yitong Li et al.

Longitudinal brain analysis is essential for understanding healthy aging and identifying pathological deviations. Longitudinal registration of sequential brain MRI underpins such analyses. However, existing methods are limited by reliance on densely sampled time series, a trade-off between accuracy and temporal smoothness, and an inability to prospectively forecast future brain states. To overcome these challenges, we introduce \emph{TimeFlow}, a learning-based framework for longitudinal brain MRI registration. TimeFlow uses a U-Net backbone with temporal conditioning to model neuroanatomy as a continuous function of age. Given only two scans from an individual, TimeFlow estimates accurate and temporally coherent deformation fields, enabling non-linear extrapolation to predict future brain states. This is achieved by our proposed inter-/extra-polation consistency constraints applied to both the deformation fields and deformed images. Remarkably, these constraints preserve temporal consistency and continuity without requiring explicit smoothness regularizers or densely sampled sequential data. Extensive experiments demonstrate that TimeFlow outperforms state-of-the-art methods in terms of both future timepoint forecasting and registration accuracy. Moreover, TimeFlow supports novel biological brain aging analyses by differentiating neurodegenerative trajectories from normal aging without requiring segmentation, thereby eliminating the need for labor-intensive annotations and mitigating segmentation inconsistency. TimeFlow offers an accurate, data-efficient, and annotation-free framework for longitudinal analysis of brain aging and chronic diseases, capable of forecasting brain changes beyond the observed study period.

Yu-Cheng Chou, Xingrui Wang, Yitong Li et al.

World engines aim to synthesize long, 3D-consistent videos that support interactive exploration of a scene under user-controlled camera motion. However, existing systems struggle under aggressive 6-DoF trajectories and complex outdoor layouts: they lose long-range geometric coherence, deviate from the target path, or collapse into overly conservative motion. To this end, we introduce Captain Safari, a pose-conditioned world engine that generates videos by retrieving from a persistent world memory. Given a camera path, our method maintains a dynamic local memory and uses a retriever to fetch pose-aligned world tokens, which then condition video generation along the trajectory. This design enables the model to maintain stable 3D structure while accurately executing challenging camera maneuvers. To evaluate this setting, we curate OpenSafari, a new in-the-wild FPV dataset containing high-dynamic drone videos with verified camera trajectories, constructed through a multi-stage geometric and kinematic validation pipeline. Across video quality, 3D consistency, and trajectory following, Captain Safari substantially outperforms state-of-the-art camera-controlled generators. It reduces MEt3R from 0.3703 to 0.3690, improves AUC@30 from 0.181 to 0.200, and yields substantially lower FVD than all camera-controlled baselines. More importantly, in a 50-participant, 5-way human study where annotators select the best result among five anonymized models, 67.6% of preferences favor our method across all axes. Our results demonstrate that pose-conditioned world memory is a powerful mechanism for long-horizon, controllable video generation and provide OpenSafari as a challenging new benchmark for future world-engine research.

Jingcong Liang, Shijun Wan, Xuehai Wu et al.

Large Reasoning Models (LRMs) have demonstrated impressive performance on complex tasks, including logical puzzle games that require deriving solutions satisfying all constraints. However, whether they can flexibly apply appropriate rules to varying conditions, particularly when faced with non-canonical game variants, remains an open question. Existing corpora focus on popular puzzles like 9x9 Sudoku, risking overfitting to canonical formats and memorization of solution patterns, which can mask deficiencies in understanding novel rules or adapting strategies to new variants. To address this, we introduce HardcoreLogic, a challenging benchmark of over 5,000 puzzles across 10 games, designed to test the robustness of LRMs on the "long-tail" of logical games. HardcoreLogic systematically transforms canonical puzzles through three dimensions: Increased Complexity (IC), Uncommon Elements (UE), and Unsolvable Puzzles (UP), reducing reliance on shortcut memorization. Evaluations on a diverse set of LRMs reveal significant performance drops, even for models achieving top scores on existing benchmarks, indicating heavy reliance on memorized stereotypes. While increased complexity is the dominant source of difficulty, models also struggle with subtle rule variations that do not necessarily increase puzzle difficulty. Our systematic error analysis on solvable and unsolvable puzzles further highlights gaps in genuine reasoning. Overall, HardcoreLogic exposes the limitations of current LRMs and establishes a benchmark for advancing high-level logical reasoning.

Yitong Li, Raoul Grasman

The modern recommender systems are facing an increasing challenge of modelling and predicting the dynamic and context-rich user preferences. Traditional collaborative filtering and content-based methods often struggle to capture the temporal patternings and evolving user intentions. While Large Language Models (LLMs) have gained gradual attention in recent years, by their strong semantic understanding and reasoning abilities, they are not inherently designed to model chronologically evolving user preference and intentions. On the other hand, for sequential models like LSTM (Long-Short-Term-Memory) which is good at capturing the temporal dynamics of user behaviour and evolving user preference over time, but still lacks a rich semantic understanding for comprehensive recommendation generation. In this study, we propose DUALRec (Dynamic User-Aware Language-based Recommender), a novel recommender that leverages the complementary strength of both models, which combines the temporal modelling abilities of LSTM networks with semantic reasoning power of the fine-tuned Large Language Models. The LSTM component will capture users evolving preference through their viewing history, while the fine-tuned LLM variants will leverage these temporal user insights to generate next movies that users might enjoy. Experimental results on MovieLens-1M dataset shows that the DUALRec model outperforms a wide range of baseline models, with comprehensive evaluation matrices of Hit Rate (HR@k), Normalized Discounted Cumulative Gain (NDCG@k), and genre similarity metrics. This research proposes a novel architecture that bridges the gap between temporal sequence modeling and semantic reasoning, and offers a promising direction for developing more intelligent and context-aware recommenders.

Yuchen Zhu, Jihong Chen, Yitong Li et al.

Structural assessment of biomolecular complexes is vital for translating molecular models into functional insights, shaping our understanding of biology and aiding drug discovery. However, current structure-based scoring functions often lack generalizability across diverse biomolecular systems. We present BioScore, a foundational scoring function that addresses key challenges -- data sparsity, cross-system representation, and task compatibility -- through a dual-scale geometric graph learning framework with tailored modules for structure assessment and affinity prediction. BioScore supports a wide range of tasks, including affinity prediction, conformation ranking, and structure-based virtual screening. Evaluated on 16 benchmarks spanning proteins, nucleic acids, small molecules, and carbohydrates, BioScore consistently outperforms or matches 70 traditional and deep learning methods. Our newly proposed PPI Benchmark further enables comprehensive evaluation of protein-protein complex scoring. BioScore demonstrates broad applicability: (1) pretraining on mixed-structure data boosts protein-protein affinity prediction by up to 40% and antigen-antibody binding correlation by over 90%; (2) cross-system generalizability enables zero- and few-shot prediction with up to 71% correlation gain; and (3) its unified representation captures chemically challenging systems such as cyclic peptides, improving affinity prediction by over 60%. BioScore establishes a robust and generalizable framework for structural assessment across complex biomolecular landscapes.