Yunfei Zhang

Yunfei Zhang, Chao Liang, Jin Gao et al.

Joint Detection and Embedding (JDE) trackers have demonstrated excellent performance in Multi-Object Tracking (MOT) tasks by incorporating the extraction of appearance features as auxiliary tasks through embedding Re-Identification task (ReID) into the detector, achieving a balance between inference speed and tracking performance. However, solving the competition between the detector and the feature extractor has always been a challenge. Meanwhile, the issue of directly embedding the ReID task into MOT has remained unresolved. The lack of high discriminability in appearance features results in their limited utility. In this paper, a new learning approach using cross-correlation to capture temporal information of objects is proposed. The feature extraction network is no longer trained solely on appearance features from each frame but learns richer motion features by utilizing feature heatmaps from consecutive frames, which addresses the challenge of inter-class feature similarity. Furthermore, our learning approach is applied to a more lightweight feature extraction network, and treat the feature matching scores as strong cues rather than auxiliary cues, with an appropriate weight calculation to reflect the compatibility between our obtained features and the MOT task. Our tracker, named TCBTrack, achieves state-of-the-art performance on multiple public benchmarks, i.e., MOT17, MOT20, and DanceTrack datasets. Specifically, on the DanceTrack test set, we achieve 56.8 HOTA, 58.1 IDF1 and 92.5 MOTA, making it the best online tracker capable of achieving real-time performance. Comparative evaluations with other trackers prove that our tracker achieves the best balance between speed, robustness and accuracy. Code is available at https://github.com/yfzhang1214/TCBTrack.

Haifeng Wang, Hua Wu, Tian Wu et al.

In this report, we introduce ERNIE 5.0, a natively autoregressive foundation model desinged for unified multimodal understanding and generation across text, image, video, and audio. All modalities are trained from scratch under a unified next-group-of-tokens prediction objective, based on an ultra-sparse mixture-of-experts (MoE) architecture with modality-agnostic expert routing. To address practical challenges in large-scale deployment under diverse resource constraints, ERNIE 5.0 adopts a novel elastic training paradigm. Within a single pre-training run, the model learns a family of sub-models with varying depths, expert capacities, and routing sparsity, enabling flexible trade-offs among performance, model size, and inference latency in memory- or time-constrained scenarios. Moreover, we systematically address the challenges of scaling reinforcement learning to unified foundation models, thereby guaranteeing efficient and stable post-training under ultra-sparse MoE architectures and diverse multimodal settings. Extensive experiments demonstrate that ERNIE 5.0 achieves strong and balanced performance across multiple modalities. To the best of our knowledge, among publicly disclosed models, ERNIE 5.0 represents the first production-scale realization of a trillion-parameter unified autoregressive model that supports both multimodal understanding and generation. To facilitate further research, we present detailed visualizations of modality-agnostic expert routing in the unified model, alongside comprehensive empirical analysis of elastic training, aiming to offer profound insights to the community.

Wentao Long, Yunfei Zhang, Chenyi Li et al.

Formal theorem-proving benchmarks enable mechanically verifiable evaluation of mathematical reasoning in large language models. However, existing benchmarks mainly focus on Olympiad-style problems and algebraic domains, leaving computational and applied mathematics underrepresented. We introduce CAM-Bench, a Lean 4 theorem-proving benchmark of 1,000 Lean proof targets in computational and applied mathematics, with coverage spanning optimization, numerical linear algebra, and numerical analysis. These problems are adapted from textbook exercises and often depend on locally introduced definitions, notation, algorithms, and elementary results. To construct CAM-Bench, we develop a dependency-recovery pipeline that reconstructs the local textbook context needed to state each problem faithfully. It then normalizes each problem into a standalone informal theorem and translates it into a Lean target. We validate the resulting formal problems through Lean compilation and semantic review, checking both formal correctness and semantic alignment with the original exercises. For each problem, we release the raw exercise, recovered context, normalized informal theorem, and final Lean target. CAM-Bench complements existing formal mathematics benchmarks by targeting applied mathematics problems that rely on textbook concepts and elementary theorems, many of which are not directly available as standard Mathlib4 lemmas. We evaluate widely used large language models and formalization agents on CAM-Bench, and analyze common failure modes in tracking local assumptions, applying elementary results, decomposing proofs, and maintaining long-horizon control in Lean.

Yunfei Zhang, Chuan Qin, Dazhong Shen et al.

During the past few decades, cognitive diagnostics modeling has attracted increasing attention in computational education communities, which is capable of quantifying the learning status and knowledge mastery levels of students. Indeed, the recent advances in neural networks have greatly enhanced the performance of traditional cognitive diagnosis models through learning the deep representations of students and exercises. Nevertheless, existing approaches often suffer from the issue of overconfidence in predicting students' mastery levels, which is primarily caused by the unavoidable noise and sparsity in realistic student-exercise interaction data, severely hindering the educational application of diagnostic feedback. To address this, in this paper, we propose a novel Reliable Cognitive Diagnosis(ReliCD) framework, which can quantify the confidence of the diagnosis feedback and is flexible for different cognitive diagnostic functions. Specifically, we first propose a Bayesian method to explicitly estimate the state uncertainty of different knowledge concepts for students, which enables the confidence quantification of diagnostic feedback. In particular, to account for potential differences, we suggest modeling individual prior distributions for the latent variables of different ability concepts using a pre-trained model. Additionally, we introduce a logical hypothesis for ranking confidence levels. Along this line, we design a novel calibration loss to optimize the confidence parameters by modeling the process of student performance prediction. Finally, extensive experiments on four real-world datasets clearly demonstrate the effectiveness of our ReliCD framework.

Jiawei Chen, Ruoxi Xu, Boxi Cao et al.

The emergence of Large Language Models (LLMs) has illuminated the potential for a general-purpose user simulator. However, existing benchmarks remain constrained to isolated scenarios, narrow action spaces, or synthetic data, failing to capture the holistic nature of authentic human behavior. To bridge this gap, we introduce OmniBehavior, the first user simulation benchmark constructed entirely from real-world data, integrating long-horizon, cross-scenario, and heterogeneous behavioral patterns into a unified framework. Based on this benchmark, we first provide empirical evidence that previous datasets with isolated scenarios suffer from tunnel vision, whereas real-world decision-making relies on long-term, cross-scenario causal chains. Extensive evaluations of state-of-the-art LLMs reveal that current models struggle to accurately simulate these complex behaviors, with performance plateauing even as context windows expand. Crucially, a systematic comparison between simulated and authentic behaviors uncovers a fundamental structural bias: LLMs tend to converge toward a positive average person, exhibiting hyper-activity, persona homogenization, and a Utopian bias. This results in the loss of individual differences and long-tail behaviors, highlighting critical directions for future high-fidelity simulation research.

Can Xie, Ruotong Pan, Xiangyu Wu et al.

Reinforcement Learning with Verifiable Rewards (RLVR) has shown significant promise for enhancing the reasoning capabilities of large language models (LLMs). However, prevailing algorithms like GRPO broadcast a uniform advantage signal across all tokens in a sequence. This coarse-grained approach overlooks the pivotal role of uncertain, high-stakes decisions during reasoning, leading to inefficient exploration and the well-documented problem of entropy collapse. To address this, we introduce UnCertainty-aware Advantage Shaping (UCAS), a model-free method that refines credit assignment by leveraging the model's internal uncertainty signals. UCAS operates in two stages: it first modulates the response-level advantage using the model's overall self-confidence, and then applies a token-level penalty based on raw logit certainty. This dual mechanism encourages exploration of high-uncertainty paths that yield correct answers while penalizing overconfident yet erroneous reasoning, effectively balancing the exploration-exploitation trade-off. Extensive experiments on five mathematical reasoning benchmarks show that UCAS significantly outperforms strong RLVR baselines across multiple model scales, including 1.5B and 7B. Our analysis confirms that UCAS not only achieves higher rewards but also promotes greater reasoning diversity and successfully mitigates entropy collapse.

Lianxin Xie, Bingbing Zheng, Wen Xue et al.

Blind face video restoration aims to restore high-fidelity details from videos subjected to complex and unknown degradations. This task poses a significant challenge of managing temporal heterogeneity while at the same time maintaining stable face attributes. In this paper, we introduce a Discrete Prior-based Temporal-Coherent content prediction transformer to address the challenge, and our model is referred to as DP-TempCoh. Specifically, we incorporate a spatial-temporal-aware content prediction module to synthesize high-quality content from discrete visual priors, conditioned on degraded video tokens. To further enhance the temporal coherence of the predicted content, a motion statistics modulation module is designed to adjust the content, based on discrete motion priors in terms of cross-frame mean and variance. As a result, the statistics of the predicted content can match with that of real videos over time. By performing extensive experiments, we verify the effectiveness of the design elements and demonstrate the superior performance of our DP-TempCoh in both synthetically and naturally degraded video restoration.

Lingzhi Zhao, Ying Cui, Zhi Liu et al.

This paper investigates adaptive streaming of one or multiple tiled 360 videos from a multi-antenna base station (BS) to one or multiple single-antenna users, respectively, in a multi-carrier wireless system. We aim to maximize the video quality while keeping rebuffering time small via encoding rate adaptation at each group of pictures (GOP) and transmission adaptation at each (transmission) slot. To capture the impact of field-of-view (FoV) prediction, we consider three cases of FoV viewing probability distributions, i.e., perfect, imperfect, and unknown FoV viewing probability distributions, and use the average total utility, worst average total utility, and worst total utility as the respective performance metrics. In the single-user scenario, we optimize the encoding rates of the tiles, encoding rates of the FoVs, and transmission beamforming vectors for all subcarriers to maximize the total utility in each case. In the multi-user scenario, we adopt rate splitting with successive decoding and optimize the encoding rates of the tiles, encoding rates of the FoVs, rates of the common and private messages, and transmission beamforming vectors for all subcarriers to maximize the total utility in each case. Then, we separate the challenging optimization problem into multiple tractable problems in each scenario. In the single-user scenario, we obtain a globally optimal solution of each problem using transformation techniques and the Karush-Kuhn-Tucker (KKT) conditions. In the multi-user scenario, we obtain a KKT point of each problem using the concave-convex procedure (CCCP). Finally, numerical results demonstrate that the proposed solutions achieve notable gains over existing schemes in all three cases. To the best of our knowledge, this is the first work revealing the impact of FoV prediction on the performance of adaptive streaming of tiled 360 videos.

Jiaoyang Yin, Yiling Xu, Hao Chen et al.

Adaptive Bit Rate (ABR) decision plays a crucial role for ensuring satisfactory Quality of Experience (QoE) in video streaming applications, in which past network statistics are mainly leveraged for future network bandwidth prediction. However, most algorithms, either rules-based or learning-driven approaches, feed throughput traces or classified traces based on traditional statistics (i.e., mean/standard deviation) to drive ABR decision, leading to compromised performances in specific scenarios. Given the diverse network connections (e.g., WiFi, cellular and wired link) from time to time, this paper thus proposes to learn the ANT (a.k.a., Accurate Network Throughput) model to characterize the full spectrum of network throughput dynamics in the past for deriving the proper network condition associated with a specific cluster of network throughput segments (NTS). Each cluster of NTS is then used to generate a dedicated ABR model, by which we wish to better capture the network dynamics for diverse connections. We have integrated the ANT model with existing reinforcement learning (RL)-based ABR decision engine, where different ABR models are applied to respond to the accurate network sensing for better rate decision. Extensive experiment results show that our approach can significantly improve the user QoE by 65.5% and 31.3% respectively, compared with the state-of-the-art Pensive and Oboe, across a wide range of network scenarios.

Donglai Zhu, Hao Chen, Hengshuai Yao et al.

The problem of action-conditioned image prediction is to predict the expected next frame given the current camera frame the robot observes and an action selected by the robot. We provide the first comparison of two recent popular models, especially for image prediction on cars. Our major finding is that action tiling encoding is the most important factor leading to the remarkable performance of the CDNA model. We present a light-weight model by action tiling encoding which has a single-decoder feedforward architecture same as [action_video_prediction_honglak]. On a real driving dataset, the CDNA model achieves ${0.3986} \times 10^{-3}$ MSE and ${0.9846}$ Structure SIMilarity (SSIM) with a network size of about {\bfseries ${12.6}$ million} parameters. With a small network of fewer than {\bfseries ${1}$ million} parameters, our new model achieves a comparable performance to CDNA at ${0.3613} \times 10^{-3}$ MSE and ${0.9633}$ SSIM. Our model requires less memory, is more computationally efficient and is advantageous to be used inside self-driving vehicles.