Yilei Wang

Yuhao Dan, Zhikai Lei, Yiyang Gu et al.

EduChat (https://www.educhat.top/) is a large-scale language model (LLM)-based chatbot system in the education domain. Its goal is to support personalized, fair, and compassionate intelligent education, serving teachers, students, and parents. Guided by theories from psychology and education, it further strengthens educational functions such as open question answering, essay assessment, Socratic teaching, and emotional support based on the existing basic LLMs. Particularly, we learn domain-specific knowledge by pre-training on the educational corpus and stimulate various skills with tool use by fine-tuning on designed system prompts and instructions. Currently, EduChat is available online as an open-source project, with its code, data, and model parameters available on platforms (e.g., GitHub https://github.com/icalk-nlp/EduChat, Hugging Face https://huggingface.co/ecnu-icalk ). We also prepare a demonstration of its capabilities online (https://vimeo.com/851004454). This initiative aims to promote research and applications of LLMs for intelligent education.

Guoqing Ma, Haoyang Huang, Kun Yan et al.

We present Step-Video-T2V, a state-of-the-art text-to-video pre-trained model with 30B parameters and the ability to generate videos up to 204 frames in length. A deep compression Variational Autoencoder, Video-VAE, is designed for video generation tasks, achieving 16x16 spatial and 8x temporal compression ratios, while maintaining exceptional video reconstruction quality. User prompts are encoded using two bilingual text encoders to handle both English and Chinese. A DiT with 3D full attention is trained using Flow Matching and is employed to denoise input noise into latent frames. A video-based DPO approach, Video-DPO, is applied to reduce artifacts and improve the visual quality of the generated videos. We also detail our training strategies and share key observations and insights. Step-Video-T2V's performance is evaluated on a novel video generation benchmark, Step-Video-T2V-Eval, demonstrating its state-of-the-art text-to-video quality when compared with both open-source and commercial engines. Additionally, we discuss the limitations of current diffusion-based model paradigm and outline future directions for video foundation models. We make both Step-Video-T2V and Step-Video-T2V-Eval available at https://github.com/stepfun-ai/Step-Video-T2V. The online version can be accessed from https://yuewen.cn/videos as well. Our goal is to accelerate the innovation of video foundation models and empower video content creators.

Ailin Huang, Boyong Wu, Bruce Wang et al.

Real-time speech interaction, serving as a fundamental interface for human-machine collaboration, holds immense potential. However, current open-source models face limitations such as high costs in voice data collection, weakness in dynamic control, and limited intelligence. To address these challenges, this paper introduces Step-Audio, the first production-ready open-source solution. Key contributions include: 1) a 130B-parameter unified speech-text multi-modal model that achieves unified understanding and generation, with the Step-Audio-Chat version open-sourced; 2) a generative speech data engine that establishes an affordable voice cloning framework and produces the open-sourced lightweight Step-Audio-TTS-3B model through distillation; 3) an instruction-driven fine control system enabling dynamic adjustments across dialects, emotions, singing, and RAP; 4) an enhanced cognitive architecture augmented with tool calling and role-playing abilities to manage complex tasks effectively. Based on our new StepEval-Audio-360 evaluation benchmark, Step-Audio achieves state-of-the-art performance in human evaluations, especially in terms of instruction following. On open-source benchmarks like LLaMA Question, shows 9.3% average performance improvement, demonstrating our commitment to advancing the development of open-source multi-modal language technologies. Our code and models are available at https://github.com/stepfun-ai/Step-Audio.

Boyong Wu, Chao Yan, Chen Hu et al.

This paper presents Step-Audio 2, an end-to-end multi-modal large language model designed for industry-strength audio understanding and speech conversation. By integrating a latent audio encoder and reasoning-centric reinforcement learning (RL), Step-Audio 2 achieves promising performance in automatic speech recognition (ASR) and audio understanding. To facilitate genuine end-to-end speech conversation, Step-Audio 2 incorporates the generation of discrete audio tokens into language modeling, significantly enhancing its responsiveness to paralinguistic information such as speaking styles and emotions. To effectively leverage the rich textual and acoustic knowledge in real-world data, Step-Audio 2 integrates retrieval-augmented generation (RAG) and is able to call external tools such as web search to mitigate hallucination and audio search to switch timbres. Trained on millions of hours of speech and audio data, Step-Audio 2 delivers intelligence and expressiveness across diverse conversational scenarios. Evaluation results demonstrate that Step-Audio 2 achieves state-of-the-art performance on various audio understanding and conversational benchmarks compared to other open-source and commercial solutions. Please visit https://github.com/stepfun-ai/Step-Audio2 for more information.

Jingyu Zhang, Yilei Wang, Lang Qian et al.

As a potential application of Vehicle-to-Everything (V2X) communication, multi-agent collaborative perception has achieved significant success in 3D object detection. While these methods have demonstrated impressive results on standard benchmarks, the robustness of such approaches in the face of complex real-world environments requires additional verification. To bridge this gap, we introduce the first comprehensive benchmark designed to evaluate the robustness of collaborative perception methods in the presence of natural corruptions typical of real-world environments. Furthermore, we propose DSRC, a robustness-enhanced collaborative perception method aiming to learn Density-insensitive and Semantic-aware collaborative Representation against Corruptions. DSRC consists of two key designs: i) a semantic-guided sparse-to-dense distillation framework, which constructs multi-view dense objects painted by ground truth bounding boxes to effectively learn density-insensitive and semantic-aware collaborative representation; ii) a feature-to-point cloud reconstruction approach to better fuse critical collaborative representation across agents. To thoroughly evaluate DSRC, we conduct extensive experiments on real-world and simulated datasets. The results demonstrate that our method outperforms SOTA collaborative perception methods in both clean and corrupted conditions. Code is available at https://github.com/Terry9a/DSRC.

Peichao Lai, Kexuan Zhang, Yi Lin et al.

Subjective Answer Grading (SAG) plays a crucial role in education, standardized testing, and automated assessment systems, particularly for evaluating short-form responses in Short Answer Scoring (SAS). However, existing approaches often produce coarse-grained scores and lack detailed reasoning. Although large language models (LLMs) have demonstrated potential as zero-shot evaluators, they remain susceptible to bias, inconsistencies with human judgment, and limited transparency in scoring decisions. To overcome these limitations, we introduce SAS-Bench, a benchmark specifically designed for LLM-based SAS tasks. SAS-Bench provides fine-grained, step-wise scoring, expert-annotated error categories, and a diverse range of question types derived from real-world subject-specific exams. This benchmark facilitates detailed evaluation of model reasoning processes and explainability. We also release an open-source dataset containing 1,030 questions and 4,109 student responses, each annotated by domain experts. Furthermore, we conduct comprehensive experiments with various LLMs, identifying major challenges in scoring science-related questions and highlighting the effectiveness of few-shot prompting in improving scoring accuracy. Our work offers valuable insights into the development of more robust, fair, and educationally meaningful LLM-based evaluation systems.

Tao Li, Peilin Li, Kui Lu et al.

The adverse drug reactions (ADRs) predicted based on the biased records in FAERS (U.S. Food and Drug Administration Adverse Event Reporting System) may mislead diagnosis online. Generally, such problems are solved by optimizing reporting odds ratio (ROR) or proportional reporting ratio (PRR). However, these methods that rely on statistical methods cannot eliminate the biased data, leading to inaccurate signal prediction. In this paper, we propose PFed-signal, a federated learning-based signal prediction model of ADR, which utilizes the Euclidean distance to eliminate the biased data from FAERS, thereby improving the accuracy of ADR prediction. Specifically, we first propose Pfed-Split, a method to split the original dataset into a split dataset based on ADR. Then we propose ADR-signal, an ADR prediction model, including a biased data identification method based on federated learning and an ADR prediction model based on Transformer. The former identifies the biased data according to the Euclidean distance and generates a clean dataset by deleting the biased data. The latter is an ADR prediction model based on Transformer trained on the clean data set. The results show that the ROR and PRR on the clean dataset are better than those of the traditional methods. Furthermore, the accuracy rate, F1 score, recall rate and AUC of PFed-Signal are 0.887, 0.890, 0.913 and 0.957 respectively, which are higher than the baselines.

Peichao Lai, Jinhui Zhuang, Kexuan Zhang et al.

Automating the conversion of UI images into web code is a critical task for front-end development and rapid prototyping. Advances in multimodal large language models (MLLMs) have made WebUI-to-Code increasingly feasible, yet existing benchmarks remain limited in data diversity and evaluation reliability. To address these issues, we present WebRenderBench, a large-scale benchmark of 45.1k webpages collected from real-world portal sites, offering greater diversity, complexity, and realism than prior benchmarks. We further propose a novel evaluation metric that measures layout and style consistency from the final rendered pages. Unlike vision-based methods that rely on costly LLM reasoning or structure-based comparisons vulnerable to noise and asymmetry, our approach enables more efficient, objective, and reliable UI quality assessment. Finally, we introduce the Automated Layout and Style Inspection Agent (ALISA), which integrates this metric into reinforcement learning as a reward signal to enhance training on crawled asymmetric webpages. Experiments show that ALISA significantly boosts generation performance, achieving state-of-the-art results across multiple metrics.

Peichao Lai, Jiaxin Gan, Feiyang Ye et al.

Sequence labeling remains a significant challenge in low-resource, domain-specific scenarios, particularly for character-dense languages like Chinese. Existing methods primarily focus on enhancing model comprehension and improving data diversity to boost performance. However, these approaches still struggle with inadequate model applicability and semantic distribution biases in domain-specific contexts. To overcome these limitations, we propose a novel framework that combines an LLM-based knowledge enhancement workflow with a span-based Knowledge Fusion for Rich and Efficient Extraction (KnowFREE) model. Our workflow employs explanation prompts to generate precise contextual interpretations of target entities, effectively mitigating semantic biases and enriching the model's contextual understanding. The KnowFREE model further integrates extension label features, enabling efficient nested entity extraction without relying on external knowledge during inference. Experiments on multiple Chinese domain-specific sequence labeling datasets demonstrate that our approach achieves state-of-the-art performance, effectively addressing the challenges posed by low-resource settings.

Zhenxing Zhang, Jun Ge, Zheng Wei et al.

The goal of feature selection is to choose the optimal subset of features for a recognition task by evaluating the importance of each feature, thereby achieving effective dimensionality reduction. Currently, proposed feature selection methods often overlook the discriminative dependencies between features and labels. To address this problem, this paper introduces a novel orthogonal regression model incorporating the area of a polygon. The model can intuitively capture the discriminative dependencies between features and labels. Additionally, this paper employs a hybrid non-monotone linear search method to efficiently tackle the non-convex optimization challenge posed by orthogonal constraints. Experimental results demonstrate that our approach not only effectively captures discriminative dependency information but also surpasses traditional methods in reducing feature dimensions and enhancing classification performance.

Qiyao Luo, Yilei Wang, Ke Yi

We present a protocol in the shuffle model of differential privacy (DP) for the \textit{frequency estimation} problem that achieves error $ω(1)\cdot O(\log n)$, almost matching the central-DP accuracy, with $1+o(1)$ messages per user. This exhibits a sharp transition phenomenon, as there is a lower bound of $Ω(n^{1/4})$ if each user is allowed to send only one message. Previously, such a result is only known when the domain size $B$ is $o(n)$. For a large domain, we also need an efficient method to identify the \textit{heavy hitters} (i.e., elements that are frequent enough). For this purpose, we design a shuffle-DP protocol that uses $o(1)$ messages per user and can identify all heavy hitters in time polylogarithmic in $B$. Finally, by combining our frequency estimation and the heavy hitter detection protocols, we show how to solve the $B$-dimensional \textit{1-sparse vector summation} problem in the high-dimensional setting $B=Ω(n)$, achieving the optimal central-DP MSE $\tilde O(n)$ with $1+o(1)$ messages per user. In addition to error and message number, our protocols improve in terms of message size and running time as well. They are also very easy to implement. The experimental results demonstrate order-of-magnitude improvement over prior work.

Qiyao Luo, Yilei Wang, Zhenghang Ren et al.

A closer integration of machine learning and relational databases has gained steam in recent years due to the fact that the training data to many ML tasks is the results of a relational query (most often, a join-select query). In a federated setting, this poses an additional challenge, that the tables are held by different parties as their private data, and the parties would like to train the model without having to use a trusted third party. Existing work has only considered the case where the training data is stored in a flat table that has been vertically partitioned, which corresponds to a simple PK-PK join. In this paper, we describe secure protocols to compute the join results of multiple tables conforming to a general foreign-key acyclic schema, and how to feed the results in secret-shared form to a secure ML toolbox. Furthermore, existing secure ML systems reveal the PKs in the join results. We strengthen the privacy protection to higher levels and achieve zero information leakage beyond the trained model. If the model itself is considered sensitive, we show how differential privacy can be incorporated into our framework to also prevent the model from breaching individuals' privacy.

Changyu Dong, Yilei Wang, Amjad Aldweesh et al.

Cloud computing has become an irreversible trend. Together comes the pressing need for verifiability, to assure the client the correctness of computation outsourced to the cloud. Existing verifiable computation techniques all have a high overhead, thus if being deployed in the clouds, would render cloud computing more expensive than the on-premises counterpart. To achieve verifiability at a reasonable cost, we leverage game theory and propose a smart contract based solution. In a nutshell, a client lets two clouds compute the same task, and uses smart contracts to stimulate tension, betrayal and distrust between the clouds, so that rational clouds will not collude and cheat. In the absence of collusion, verification of correctness can be done easily by crosschecking the results from the two clouds. We provide a formal analysis of the games induced by the contracts, and prove that the contracts will be effective under certain reasonable assumptions. By resorting to game theory and smart contracts, we are able to avoid heavy cryptographic protocols. The client only needs to pay two clouds to compute in the clear, and a small transaction fee to use the smart contracts. We also conducted a feasibility study that involves implementing the contracts in Solidity and running them on the official Ethereum network.