Inverse Protein Folding (IPF) is a critical subtask in the field of protein design, aiming to engineer amino acid sequences capable of folding correctly into a specified three-dimensional (3D) conformation. Although substantial progress has been achieved in recent years, existing methods generally rely on either backbone coordinates or molecular surface features alone, which restricts their ability to fully capture the complex chemical and geometric constraints necessary for precise sequence prediction. To address this limitation, we present DS-ProGen, a dual-structure deep language model for functional protein design, which integrates both backbone geometry and surface-level representations. By incorporating backbone coordinates as well as surface chemical and geometric descriptors into a next-amino-acid prediction paradigm, DS-ProGen is able to generate functionally relevant and structurally stable sequences while satisfying both global and local conformational constraints. On the PRIDE dataset, DS-ProGen attains the current state-of-the-art recovery rate of 61.47%, demonstrating the synergistic advantage of multi-modal structural encoding in protein design. Furthermore, DS-ProGen excels in predicting interactions with a variety of biological partners, including ligands, ions, and RNA, confirming its robust functional retention capabilities.

Recent advances in video understanding have been driven by MLLMs. But these MLLMs are good at analyzing short videos, while suffering from difficulties in understanding videos with a longer context. To address this difficulty, several agent paradigms have recently been proposed, using MLLMs as agents for retrieving extra contextual knowledge in a long video. However, most existing agents ignore the key fact that a long video is composed with multiple shots, i.e., to answer the user question from a long video, it is critical to deeply understand its relevant shots like human. Without such insight, these agents often mistakenly find redundant even noisy temporal context, restricting their capacity for long video understanding. To fill this gap, we propose VideoChat-A1, a novel long video agent paradigm. Different from the previous works, our VideoChat-A1 can deeply think with long videos, via a distinct chain-of-shot reasoning paradigm. More specifically, it can progressively select the relevant shots of user question, and look into these shots in a coarse-to-fine partition. By multi-modal reasoning along the shot chain, VideoChat-A1 can effectively mimic step-by-step human thinking process, allowing the interactive discovery of preferable temporal context for thoughtful understanding in long videos. Extensive experiments show that, VideoChat-A1 achieves the state-of-the-art performance on the mainstream long video QA benchmarks, e.g., it achieves 77.0 on VideoMME(w/ subs) and 70.1 on EgoSchema, outperforming its strong baselines (e.g., InternVL2.5-8B and InternVideo2.5-8B), by up to 10.1% and 6.2%. Compared to leading closed-source GPT-4o and Gemini 1.5 Pro, VideoChat-A1 offers competitive accuracy, but only with 7% input frames and 12% inference time on average.

Andrew Zhuoer Feng, Cunxiang Wang, Yu Luo, Lin Fan, Irene Zhou, Zikang Wang, Xiaotao Gu, Jie Tang, Hongning Wang, Minlie Huang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

The evolution of modern communication and the Internet has enhanced the accessibility, editability, and dissemination speed of the digital images, intensifying the need for digital image copyright protection. Digital watermarking technique is widely applied to the copyright protection of digital images. Compared with binary watermarking, multi-value watermarking can convey richer copyright information while offering higher robustness. However, existing multi-value watermarking algorithms suffer from high computational complexity, which makes it difficult to meet the ever-increasing demands for real-time and high-speed images or videos processing tasks. In this paper, we propose the TEA-M, a tiny and efficient multi-value watermarking architecture. By applying a multiplier-free approximate discrete cosine transform (DCT) transform to the Z channel and leveraging two novel multi-value watermarking strategies based on remainder adjustment, TEA-M achieves up to 49× higher energy efficiency and 310× higher area efficiency while maintaining high throughput compared with the latest approaches. For the FPGA implementation, TEA-M can achieve 5007 frames/s at its highest operating frequency of 328.299 MHz, with a peak energy efficiency of 4.146 × 105Mbps/W. For the ASIC implementation, TEA-M achieves 7629 frames/s at 500 MHz, with a peak energy efficiency of 5.36×106Mbps/W and an area efficiency of 2.1×106Mbps/mm2. In practical applications, TEA-M further offers users diverse watermarking configuration options to meet the needs of various scenarios.

Knowledge graph completion (KGC) addresses the issue of incomplete knowledge graphs by inferring missing triples, which is crucial for information retrieval, question answering, and recommender systems. Existing KGC methods generally focus on either exploiting the graph’s structural topology or leveraging the semantic information from entity descriptions. However, existing approaches often overlook the synergy between structure information and semantic information. To address the existing shortcomings, we proposeStrucSem, a novel model that leverages both structural and semantic information to boost KGC performance. Our model: 1) encodes the graph’s structural information by aggregating neighborhood data around the query entity using an attention mechanism; and 2) combines this with the encoding of textual descriptions, facilitating the integration of both types of information. Additionally, we extend contrastive learning to incorporate multiple positive samples, improving the model’s ability to represent diverse relational patterns. Our approach significantly enhances KGC performance, as demonstrated through extensive evaluations on standard benchmark datasets. The results highlight the superiority of combining structural and semantic information, offering new insights into improving KGC tasks.

Jiyue Jiang, Yitao Xu, Zikang Wang, Yihan Ye, Yanruisheng Shao, Yuheng Shan, Jiuming Wang, Xiaodan Fan, Jiao Yuan, Yu Li. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Existing MLLMs encounter significant challenges in modeling the temporal context within long videos. Currently, mainstream Agent-based methods use external tools to assist a single MLLM in answering long video questions. Despite such tool-based support, a solitary MLLM still offers only a partial understanding of long videos, resulting in limited performance. In order to better address long video tasks, we introduce LVAgent, the first framework enabling multi-round dynamic collaboration of MLLM agents in long video understanding. Our method consists of four key steps: 1) Selection: We pre-select appropriate agents from the model library to form optimal agent teams based on different tasks. 2) Perception: We design an effective retrieval scheme for long videos to improve the coverage of critical temporal segments while maintaining computational efficiency. 3) Action: Agents answer long video questions and exchange reasons. 4) Reflection: We evaluate each agent's performance in each round of discussion and optimize the agent team for dynamic collaboration. The agents iteratively refine their answers by multi-round dynamical collaboration of MLLM agents. LVAgent is the first agent system method that outperforms all closed-source models (like GPT-4o) and open-source models (like InternVL-2.5 and Qwen2-VL) in the long video understanding tasks. Our LVAgent achieves an accuracy of 80\% on four mainstream long video understanding tasks. Notably, LVAgent improves accuracy by 13.3\% on LongVideoBench. Code is available at https://github.com/64327069/LVAgent.

Autonomous systems have advanced significantly, but challenges persist in accident-prone environments where robust decision-making is crucial. A single vehicle’s limited sensor range and obstructed views increase the likelihood of accidents. Multi-vehicle connected systems and multi-modal approaches, leveraging RGB images and LiDAR point clouds, have emerged as promising solutions. However, existing methods often assume the availability of all data modalities and connected vehicles during both training and testing, which is impractical due to potential sensor failures or missing connected vehicles. To address these challenges, we introduce a novel framework MMCD (Multi-Modal Collaborative Decision-making) for connected autonomy. Our framework fuses multi-modal observations from ego and collaborative vehicles to enhance decision-making under challenging conditions. To ensure robust performance when certain data modalities are unavailable during testing, we propose an approach based on cross-modal knowledge distillation with a teacher-student model structure. The teacher model is trained with multiple data modalities, while the student model is designed to operate effectively with reduced modalities. In experiments on connected autonomous driving with ground vehicles and aerial-ground vehicles collaboration, our method improves driving safety by up to 20.7%, surpassing the best-existing baseline in detecting potential accidents and making safe driving decisions. More information can be found on our website https://ruiiu.github.io/mmcd.

In this paper, we describe veDB-HTAP, a highly integrated, efficient, and adaptive HTAP system recently built in ByteDance. veDB-HTAP adopts a highly integrated system architecture by leveraging the Secondary Engine mechanism provided by MySQL and provides a seamless query processing experience across OLTP and OLAP engines. In addition, we introduce a cost-based and machine-learning-based smart query router that significantly outperforms the rule-based query router used in ByteHTAP, a precursor of veDB-HTAP. A key design principle of veDB-HTAP is the collaboration and adaptability of major system components, including query planning, query execution, and unified storage. Our adaptive query execution can be classified into two categories: 1) adaptive execution that dynamically collects and utilizes runtime statistics for better query performance; 2) utilizing runtime resource information to achieve a high quality of service even under heavy workloads. The experiments show that veDB-HTAP can achieve more than 3× speedup for TPC-H while consuming only one-third of the resources compared to ByteHTAP.

Knowledge reasoning has been widely researched in knowledge graphs (KGs), but there has been relatively less research on hyper-relational KGs, which also plays an important role in downstream tasks. Existing reasoning methods on hyper-relational KGs are based on representation learning. Though this approach is effective, it lacks interpretability and ignores the graph structure information. In this paper, we make the first attempt at symbolic reasoning on hyper-relational KGs. We introduce rule extraction methods based on both individual facts and paths, and propose a rule-based symbolic reasoning approach, HyperPath. This approach is simple and interpretable, it can serve as a baseline model for symbolic reasoning in hyper-relational KGs. We provide experimental results on almost all datasets, including five large-scale datasets and seven sub-datasets of them. Experiments show that the expressive power of the proposed model is similar to simple neural networks like convolutional networks, but not as advanced as more complex networks such as Transformer and graph convolutional networks, which is consistent with the performance of symbolic methods on KGs. Furthermore, we also analyze the impact of rule length and hyperparameters on the model's performance, which can provide insights for future research in hypergraph symbolic reasoning.

In recent years, few-shot knowledge graph completion (FKGC) emerged as a prominent research problem, focused on utilizing a limited number of reference entity pairs to complete triples with unseen relations. Recent studies have attempted addressing this problem by modeling interactions between head and tail entities. However, existing FKGC methods represent semantics predominantly based on the neighborhood information of entities in the knowledge graph, thus can only infer the hidden and unobserved relations within the knowledge graph, limiting their reasoning capabilities. To overcome these limitations, we introduce text descriptions to FKGC and propose BERT-FKGC, a model capable of learning the integrated distribution of both the entity text descriptions and neighborhood information. By using a gating network that allows the model to dynamically select weights, our method can flexibly combine neighborhood information and textual descriptions. Besides addressing the prediction of unseen relations, our method is also capable of representing unseen entities. To validate the effectiveness of our model, we introduce a new dataset, FB15K-237-One, which includes textual descriptions for entities. We conduct extensive experiments on the FB15K-237-One dataset to validate the superiority of BERTFKGC.

Few-shot, especially one-shot learning is a prominent research area in the field of knowledge graphs (KGs), aiming to utilize a limited number of triples with unseen relations as reference information for inferring missing knowledge. Recent research focuses on improving the semantic representation of entity pairs using interactions between their head and tail entities. However, this method only considers the reference information as the measurement criterion without taking into account the potential impact of it on the reasoning process of the model. In this paper, we propose a novel method that utilizes factual information interactions. Firstly, we learn static representations of entities based on their neighborhood information. Subsequently, we learn relation adaptive representations by incorporating the reference information. This interactive modeling strengthens the association between entity representations and task relations while suppressing irrelevant relations. Extensive experiments demonstrate that our model outperforms state-of-the-art methods on two public datasets. Remarkably, on the NELL-One dataset for one-shot link prediction, our model achieves an improvement of 11.8% in MRR compared to the best baseline model.

Traditional long-tailed learning methods often perform poorly when dealing with inconsistencies between training and test data distributions, and they cannot flexibly adapt to different user preferences for trade-offs between head and tail classes. To address this issue, we propose a novel long-tailed learning paradigm that aims to tackle distribution shift in real-world scenarios and accommodate different user preferences for the trade-off between head and tail classes. We generate a set of diverse expert models via hypernetworks to cover all possible distribution scenarios, and optimize the model ensemble to adapt to any test distribution. Crucially, in any distribution scenario, we can flexibly output a dedicated model solution that matches the user's preference. Extensive experiments demonstrate that our method not only achieves higher performance ceilings but also effectively overcomes distribution shift while allowing controllable adjustments according to user preferences. We provide new insights and a paradigm for the long-tailed learning problem, greatly expanding its applicability in practical scenarios. The code can be found here: https://github.com/DataLab-atom/PRL.

Decentralized trading schemes involving energy prosumers have prevailed in recent years. Such schemes provide a pathway for increased energy efficiency and can be enhanced by the use of blockchain technology to address security concerns in decentralized trading. To improve transaction security and privacy protection while ensuring desirable social governance, this article proposes a novel two-stage blockchain-based operation and trading mechanism to enhance energy hubs connected with integrated energy systems (IESs). This mechanism includes multienergy aggregators (MAGs) that use a consortium blockchain and its enabled proof-of-work (PoW) to transfer and audit transaction records, with social governance principles for guiding prosumers’ decision-making in the peer-to-peer (P2P) transaction management process. The uncertain nature of renewable generation and load demand are adequately modeled in the two-stage Wasserstein-based distributionally robust optimization (DRO). The practicality of the proposed mechanism is illustrated by several case studies that jointly show its ability to handle an increased renewable generation capacity, achieve a 16.7% saving in the audit cost, and facilitate 2.4% more P2P interactions. Overall, the proposed two-stage blockchain-based trading mechanism provides a practical trading scheme and can reduce redundant trading amounts by 6.5%, leading to a further reduction of the overall operation cost. Compared to the state-of-the-art benchmark methods, our mechanism exhibits significant operation cost reduction and ensures social governance and transaction security for IES and energy hubs.

Script event prediction aims to infer subsequent events given an incomplete script. It requires a deep understanding of events, and can provide support for a variety of tasks. Existing models rarely consider the relational knowledge between events, they regard scripts as sequences or graphs, which cannot capture the relational information between events and the semantic information of script sequences jointly. To address this issue, we propose a new script form, relational event chain, that combines event chains and relational graphs. We also introduce a new model, relational-transformer, to learn embeddings based on this new script form. In particular, we first extract the relationship between events from an event knowledge graph to formalize scripts as relational event chains, then use the relational-transformer to calculate the likelihood of different candidate events, where the model learns event embeddings that encode both semantic and relational knowledge by combining transformers and graph neural networks (GNNs). Experimental results on both one-step inference and multistep inference tasks show that our model can outperform existing baselines, indicating the validity of encoding relational knowledge into event embeddings. The influence of using different model structures and different types of relational knowledge is analyzed as well.

Ensemble clustering combines the results of multiple individual clustering methods for better results. Basically, all available clustering methods can be combined to produce final clusters. However, selecting a subset of optimal methods can reduce the complexity and increase the efficiency of ensemble clustering methods. This article examines the problem of selecting individual clustering methods to produce an ensemble hierarchical clustering method. Hierarchical clustering is a technique for grouping data at different scales by creating dendrograms. The aim is to select a subset of individual hierarchical clustering methods considering diversity and quality that can create an ensemble clustering method with minimal complexity. The proposed method consists of three main phases. The selection of a subset of individual hierarchical clustering methods is done in the first phase. In the second phase, the results of the selected clusters are re-clustered to create super-clusters. Super-clusters can combine clustering knowledge of different methods into one clustering form. Finally, the final clusters are formed by assigning each sample to a super-cluster with the shortest distance in the third phase. Experimental results on several datasets from the University of California Irvine (UCI) repository show that the proposed method performs better than the state-of-the-art algorithms.

In recent years, at ByteDance, we have started seeing more and more business scenarios that require performing real-time data serving besides complex Ad Hoc analysis over large amounts of freshly imported data. The serving workload requires performing complex queries over massive newly added data items with minimal delay. These systems are often used in mission-critical scenarios, whereas traditional OLAP systems cannot handle such use cases. To work around the problem, ByteDance products often have to use multiple systems together in production, forcing the same data to be ETLed into multiple systems, causing data consistency problems, wasting resources, and increasing learning and maintenance costs. To solve the above problem, we built a single Hybrid Serving and Analytical Processing (HSAP) system to handle both workload types. HSAP is still in its early stage, and very few systems are yet on the market. This paper demonstrates how to build Krypton, a competitive cloud-native HSAP system that provides both excellent elasticity and query performance by utilizing many previously known query processing techniques, a hierarchical cache with persistent memory, and a native columnar storage format. Krypton can support high data freshness, high data ingestion rates, and strong data consistency. We also discuss lessons and best practices we learned in developing and operating Krypton in production.

Representation learning is a fundamental task in knowledge graph-related research and applications. Most existing approaches learn representations for entities and relations only based on static facts, where temporal information has been ignored completely. This paper aims to learn time-aware representations for entities and relations in knowledge graphs. Based on how temporal information affects the learned embeddings, we propose three assumptions and build three different models, BTS, ETS, and RTS, respectively. In these models, we build two separate embedding spaces for entities and relations, the standard translation condition is checked after projecting embedding vectors between these spaces by model-specific transformations. As to the performance, the proposed RTS model achieves state-of-the-art results in three experiments conducted on two datasets: YAGO11k and Wikidata12k, which validates the effectiveness of our model. Comparing the results of all three models, we find that relation embeddings are time-sensitive and form natural ordering, while the effects of time on entity embeddings can be safely ignored for translation-based methods. Experiments also show that our findings can be used to simplify other existing models like HyTE.

Natural Language Inference (NLI) is a vital task in natural language processing.It aims to identify the logical relationship between two sentences.Most of the existing approaches make such inference based on semantic knowledge obtained through training corpus.The adoption of background knowledge is rarely seen or limited to a few specific types.In this paper, we propose a novel Knowledge Graph-enhanced NLI (KGNLI) model to leverage the usage of background knowledge stored in knowledge graphs in the field of NLI.KGNLI model consists of three components: a semantic-relation representation module, a knowledge-relation representation module, and a label prediction module.Different from previous methods, various kinds of background knowledge can be flexibly combined in the proposed KGNLI model.Experiments on four benchmarks, SNLI, MultiNLI, SciTail, and BNLI, validate the effectiveness of our model.

Knowledge graph completion, one of the most important research questions in knowledge graphs, aims at predicting missing links in a given graph. Current mainstream approaches adopt high-quality embeddings of entities and relations of the graph to improve their performances. However, it is not easy to devise a universal embedding learner that can fit various scenarios. In this paper, we propose a general-purpose framework which can be employed to improve the performance of knowledge graph completion. Specifically, given an arbitrary knowledge graph completion model, we first run the original model to get a ranked entity list. Then, we combine the query and the top ranked entities with attention mechanism, re-rank all these entities by feeding the combined vector into a neural network. The proposed re-ranking phase can be conveniently added to a variety of models to improve their performance without substantial modification. We conduct experiments on four datasets: WN18, FB15k, WN18RR, and FB15k-237. We choose TransE, TransH, TransD, DistMult, and ANALOGY as base models. Experiments on these datasets and models validate the effectiveness of the proposed re-ranking framework. We further explore the influence of the number of top ranked entities used in the re-ranking phase. We also test other attention mechanism to determine the most effective one, and found that vanilla attention mechanism can balance accuracy and complexity.

Knowledge graph, or knowledge base, plays an important role in a variety of applications in the field of artificial intelligence. In both research and application of knowledge graph, knowledge representation learning is one of the fundamental tasks. Existing representation learning approaches are mainly based on structural knowledge between entities and relations, while knowledge among entities per se is largely ignored. Though a few approaches integrated entity knowledge while learning representations, these methods lack the flexibility to apply to multimodalities. To tackle this problem, in this paper, we propose a new representation learning method, TransAE, by combining multimodal autoencoder with TransE model, where TransE is a simple and effective representation learning method for knowledge graphs. In TransAE, the hidden layer of autoencoder is used as the representation of entities in the TransE model, thus it encodes not only the structural knowledge, but also the multimodal knowledge, such as visual and textural knowledge, into the final representation. Compared with traditional methods based on only structural knowledge, TransAE can significantly improve the performance in the sense of link prediction and triplet classification. Also, TransAE has the ability to learn representations for entities out of knowledge base in zero-shot. Experiments on various tasks demonstrate the effectiveness of our proposed TransAE method.

Knowledge graph plays an important role in detection, prediction, early warning, and other security related applications. A fundamental task in applying knowledge graph is the so-called multi-hop reasoning, which focuses on inferring new relations between entities. In this paper, we introduce attention mechanism to the classic compositional method. After finding reasoning paths between entities, we aggregate these paths' embeddings into one according to their attentions, and infer the relation of entities based on the combined embedding. Two experiments on NELL-995 dataset, fact prediction and link prediction, validated that our method outperforms all baselines.

Vehicular Swarm Network (VSN) technology enables real time inter-vehicular communications for broadcasting all kinds of data collection, including emergency video with high data rate. A challenge in VSN broadcasting is to achieve a high throughput rate while at the same time assure the delivery of video packet flows to all the vehicles traveling over a highway segment from the accident vehicle. This paper analyzed the throughput- distance relationship model of 802.11p standard in multi-relay VSN with Vehicle Backbone Sub-Network (VBSN). Also, this paper proposed Enhanced Throughput Capacity (ETC) scheme; using GPS, vehicles that positioned close to those throughput-optimal positions are elected as VSBN relays. The aim of this paper is two-fold: a. We analyze the property of optimal position for enhancing the broadcast throughput. b. We demonstrate the performance offered by this ETC scheme; it excludes additional VBSN rebuilt process, and has robustness of high throughput in different traffic density and vehicle speed.

This paper presents a fuzzy logic based terrain identification method using multi-sensor fusion for powered prosthesis control. Five locomotion features including rising time of ground reaction force, sequence of foot strike on ground, foot inclination angle during stance, shank inclination angle at toe-off and maximal shank inclination angle during swing are selected to identify different terrains. These features are measured by fusion of two gyroscopes, two accelerometers, two force sensitive resistors and a timer. Based on the features, a fuzzy logic identification method is developed to identify level-ground, stair ascent, stair descent, upslope and downslope online in real time. Average identification accuracy higher than 97.5% is obtained in experiments of five able-bodied subjects and a transtibial amputee. Continuous identification results show the prospect of using the proposed method to realize real-time terrain identification of powered prostheses.