arXiv subjects

Lan Wang

Publications and source records attributed to Lan Wang.

At least 37 records · Page 2Linked to original sources

arXiv2026

SeedProteo: Accurate De Novo All-Atom Design of Protein Binders

We present SeedProteo, a diffusion-based model for de novo all-atom protein design. We demonstrate how to repurpose a cutting-edge folding architecture into a powerful generative design framework by effectively integrating self-conditioning features. Extensive benchmarks highlight the model's capabilities across two distinct tasks: in unconditional generation, SeedProteo exhibits superior length generalization and structural diversity, maintaining robustness for long sequences and complex topologies; in binder design, it achieves state-of-the-art performance among open-source methods, attaining the highest in-silico design success rates, structural diversity and novelty. Finally, we validate SeedProteo through wet-lab assays on two therapeutic targets, achieving hit rates of 70%-80% and picomolar-level binding affinities, establishing leading results. To facilitate community adoption, we provide public access to SeedProteo via a webserver (https://seedfold.io/proteinDesign).

q-bio.BM↗

arXiv2026

Double Fairness Policy Learning: Integrating Action Fairness and Outcome Fairness in Decision-making

Fairness is a central pillar of trustworthy machine learning, especially in domains where accuracy- or profit-driven optimization is insufficient. While most fairness research focuses on supervised learning, fairness in policy learning remains less explored. Because policy learning is interventional, it induces two distinct fairness targets: action fairness (equitable action assignments) and outcome fairness (equitable downstream consequences). Crucially, equalizing actions does not generally equalize outcomes when groups face different constraints or respond differently to the same action. We propose a novel double fairness learning (DFL) framework that explicitly manages the trade-off among three objectives: action fairness, outcome fairness, and value maximization. We integrate fairness directly into a multi-objective optimization problem for policy learning and employ a lexicographic weighted Tchebyshev method that recovers Pareto solutions beyond convex settings, with theoretical guarantees on the regret bounds. Our framework is flexible and accommodates various commonly used fairness notions. Extensive simulations demonstrate improved performance relative to competing methods. In applications to a motor third-party liability insurance dataset and an entrepreneurship training dataset, DFL substantially improves both action and outcome fairness while incurring only a modest reduction in overall value.

stat.ML↗

arXiv2026

Looping Back to Move Forward: Recursive Transformers for Efficient and Flexible Large Multimodal Models

Large Multimodal Models (LMMs) have achieved remarkable success in vision-language tasks, yet their vast parameter counts are often underutilized during both training and inference. In this work, we embrace the idea of looping back to move forward: reusing model parameters through recursive refinement to extract stronger multimodal representations without increasing model size. We propose RecursiveVLM, a recursive Transformer architecture tailored for LMMs. Two key innovations enable effective looping: (i) a Recursive Connector that aligns features across recursion steps by fusing intermediate-layer hidden states and applying modality-specific projections, respecting the distinct statistical structures of vision and language tokens; (ii) a Monotonic Recursion Loss that supervises every step and guarantees performance improves monotonically with recursion depth. This design transforms recursion into an on-demand refinement mechanism: delivering strong results with few loops on resource-constrained devices and progressively improving outputs when more computation resources are available. Experiments show consistent gains of +3% over standard Transformers and +7% over vanilla recursive baselines, demonstrating that strategic looping is a powerful path toward efficient, deployment-adaptive LMMs.

cs.LG↗

arXiv2026

The impact of cosmic filaments on the abundance of satellite galaxies

The impact of cosmic web environments on galaxy properties plays a critical role in understanding galaxy formation. Using the state-of-the-art cosmological simulation IllustrisTNG, we investigate how satellite galaxy abundance differs between filaments and the field, with filaments identified using the DisPerSE algorithm. When filaments are identified using galaxies as tracers, we find that, across all magnitude bins, central galaxies in filaments tend to host more satellite galaxies than their counterparts in the field, in qualitative agreement with observational results from the Sloan Digital Sky Survey. The average ratios between satellite luminosity functions in filaments and the field are $3.49$, $2.61$, and $1.90$ in the central galaxy $r$-band magnitude bins of $M_{r, {\rm cen}} \sim -22$, $-21$, and $-20$, respectively. We show that much of this excess can be attributed to the higher host halo masses of galaxies in filaments. After resampling central galaxies in both environments to match the halo mass distributions within each magnitude bin, the satellite abundance enhancement in filaments is reduced by up to $79 \%$. Additionally, the choice of tracers used to identify filaments introduces a significant bias: when filaments are identified using the dark matter density field, the environmental difference in satellite abundance is reduced by more than $70 \%$; after further resampling in both magnitude and halo mass, the difference is further suppressed by another $\sim 60$--$95 \%$. Our results highlight the importance of halo mass differences and tracer choice biases when interpreting and understanding the impact of environment on satellite galaxy properties.

astro-ph.GA↗

arXiv2025

SRM at 30: Lessons from Early Data-Centric Networking and Their Impact on Named Data Networking

A 1995 SIGCOMM paper, "A Reliable Multicast Framework for Light-weight Sessions and Application-Level Framing", commonly known as SRM, explored a fundamentally new approach to reliable multiparty data delivery. Rather than adapting established sender-driven reliable unicast mechanisms to multicast, as most contemporaneous proposals did, SRM introduced a data-centric model in which data receivers recover losses by explicitly requesting missing data. Thirty years later, we revisit the SRM framework, examining the challenges it faced, the lessons learned, and its influence on the later development of Named Data Networking (NDN). Experimentations with SRM revealed a fundamental semantic mismatch between its data-centric framework and IP's address-based delivery; while the application layer named data, the network layer remained 'blind' to those names, resulting in inefficient loss recovery. NDN resolves this architectural friction by aligning network delivery with the data-retrieval model and by securing data directly rather than securing communication channels. This retrospective highlights how early insights from SRM informed key design decisions in NDN and illustrates how NDN's design emerged from the cumulative insights gained over decades of networking research and development.

cs.NI↗

arXiv2025

SeedFold: Scaling Biomolecular Structure Prediction

Highly accurate biomolecular structure prediction is a key component of developing biomolecular foundation models, and one of the most critical aspects of building foundation models is identifying the recipes for scaling the model. In this work, we present SeedFold, a folding model that successfully scales up the model capacity. Our contributions are threefold: first, we identify an effective width-scaling strategy for the Pairformer to increase representation capacity; second, we introduce a novel linear triangular attention that reduces computational complexity to enable efficient scaling; finally, we construct a large-scale distillation dataset to substantially enlarge the training set. Experiments on FoldBench show that SeedFold outperforms AlphaFold3 on most protein-related tasks.

q-bio.BM↗

arXiv2025

Interface-Controlled Antiferromagnetic Tunnel Junctions based on a metallic van der Waals A-type Antiferromagnet

Magnetic tunnel junctions (MTJs) are crucial components in high-performance spintronic devices. Traditional MTJs rely on ferromagnetic (FM) materials but significant improvements in speed and packing density could be enabled by exploiting antiferromagnetic (AFM) compounds instead. Here, we report all-collinear AFM tunnel junctions (AFMTJs) fabricated with van der Waals A-type AFM metal (Fe0.6Co0.4)5GeTe2 (FCGT) electrodes and nonmagnetic semiconducting WSe2 tunnel barriers. The AFMTJ heterostructure device achieves a tunneling magnetoresistance (TMR) ratio of up to 75% in response to magnetic field switching. Our results demonstrate that the TMR exclusively emerges in the AFM state of FCGT, rather than during the AFM-to-FM transition. By engineering FCGT electrodes with either even- or odd-layer configurations, volatile or non-volatile TMR could be selected, consistent with an entirely interfacial effect. TMR in the even-layer devices arose by Néel vector switching. In the odd-layer devices, TMR stemmed from interfacial spin-flipping. Experimental and theoretical analyses reveal a new TMR mechanism associated with interface-driven spin-polarized transport, despite the spin-independent nature of bulk FCGT. Our work demonstrates that collinear AFMTJs can provide comparable performance to conventional MTJs and introduces a new paradigm for AFM spintronics, in which the spin-dependent properties of AFM interfaces are harnessed.

cond-mat.mes-hall↗

arXiv2025

X-Talk: On the Underestimated Potential of Modular Speech-to-Speech Dialogue System

We present X-Talk, an open-source framework that champions a decoupled, modular design for LLM-driven speech-to-speech (S2S) systems. While the dominant trend favors end-to-end (E2E) modeling to optimize information flow, these "omni-models" often struggle to balance the competing objectives of complex speech tasks within a single network. X-Talk challenges this paradigm by demonstrating that a systematically optimized cascaded pipeline can achieve sub-second latency without sacrificing modular flexibility. Our framework seamlessly integrates specialized front-end components (e.g., VAD, speech enhancement) and diverse understanding models (e.g., ASR, emotion, and environmental sound analysis) with LLM capabilities like retrieval-augmented generation (RAG) and tool use. By revitalizing the cascaded approach, X-Talk highlights the underestimated potential of modular S2S systems and provides a robust foundation for future research and applications.

cs.SD↗

arXiv2025

OrdMoE: Preference Alignment via Hierarchical Expert Group Ranking in Multimodal Mixture-of-Experts LLMs

Preference learning has recently emerged as a pivotal strategy for post-training alignment of Multimodal Large Language Models (MLLMs). However, existing approaches predominantly rely on external human-annotated preference data, which is costly and labor-intensive to collect. In this work, we propose OrdMoE, a novel preference alignment framework that bypasses the reliance on external human preferences entirely by leveraging intrinsic signals within Mixture-of-Experts (MoE) architectures. Specifically, we observe that the router's expert selection scores implicitly encode a quality-aware ranking of responses (i.e. higher-scoring experts consistently generate higher-quality outputs). Building on this insight, OrdMoE constructs an internal preference hierarchy by grouping experts into ranked tiers based on their per-token routing scores and activating each tier separately to produce a sequence of responses with increasing quality. This yields a zero-cost, self-supervised preference ordering over generated responses, which can be directly optimized using standard preference learning objectives. Extensive experiments across multiple multimodal benchmarks demnstrate that OrdMoE significantly enhances both alignment and overall performance of multimodal Mixture-of-Experts LLMs, achieving competitive results without requiring any human-annotated preference data.

cs.LG↗

arXiv2025

Causally-Grounded Dual-Path Attention Intervention for Object Hallucination Mitigation in LVLMs

Object hallucination remains a critical challenge in Large Vision-Language Models (LVLMs), where models generate content inconsistent with visual inputs. Existing language-decoder based mitigation approaches often regulate visual or textual attention independently, overlooking their interaction as two key causal factors. To address this, we propose Owl (Bi-mOdal attention reWeighting for Layer-wise hallucination mitigation), a causally-grounded framework that models hallucination process via a structural causal graph, treating decomposed visual and textual attentions as mediators. We introduce VTACR (Visual-to-Textual Attention Contribution Ratio), a novel metric that quantifies the modality contribution imbalance during decoding. Our analysis reveals that hallucinations frequently occur in low-VTACR scenarios, where textual priors dominate and visual grounding is weakened. To mitigate this, we design a fine-grained attention intervention mechanism that dynamically adjusts token- and layer-wise attention guided by VTACR signals. Finally, we propose a dual-path contrastive decoding strategy: one path emphasizes visually grounded predictions, while the other amplifies hallucinated ones -- letting visual truth shine and hallucination collapse. Experimental results on the POPE and CHAIR benchmarks show that Owl achieves significant hallucination reduction, setting a new SOTA in faithfulness while preserving vision-language understanding capability. Our code is available at https://github.com/CikZ2023/OWL

cs.CV↗

arXiv2025

UTI-LLM: A Personalized Articulatory-Speech Therapy Assistance System Based on Multimodal Large Language Model

Speech therapy is essential for rehabilitating speech disorders caused by neurological impairments such as stroke. However, traditional manual and computer-assisted systems are limited in real-time accessibility and articulatory motion feedback. Recent advances in multimodal large language models (MLLMs) have demonstrated significant potential in healthcare, especially through their adaptive assessment and therapeutic feedback capabilities. Nevertheless, challenges including insufficient acquisition and fusion of articulatory information, inadequate parsing of articulatory organ motion trajectories, and the scarcity of domain-specific datasets hinder the application of MLLMs in speech therapy. To address these limitations, we propose an MLLM-based speech rehabilitation assistance system that leverages ultrasound tongue imaging and speech signals to deliver precise, interactive articulatory feedback. We construct a high-quality domain-specific dataset comprising ultrasound-speech dialogue pairs. This dataset facilitates fine-tuning to enhance the model's clinical adaptability. Furthermore, our method develops spatiotemporal fusion training strategy of ultrasound videos and speech signals, enabling fine-grained articulatory impairment analysis and ultimately generating actionable feedback. Experimental results demonstrate the effectiveness of our model in articulatory analysis and clinical assessment.

cs.SD↗

arXiv2025

Magnitude and Phase-based Feature Fusion Using Co-attention Mechanism for Speaker recognition

Phase-based features related to vocal source characteristics can be incorporated into magnitude-based speaker recognition systems to improve the system performance. However, traditional feature-level fusion methods typically ignore the unique contributions of speaker semantics in the magnitude and phase domains. To address this issue, this paper proposed a feature-level fusion framework using the co-attention mechanism for speaker recognition. The framework consists of two separate sub-networks for the magnitude and phase domains respectively. Then, the intermediate high-level outputs of both domains are fused by the co-attention mechanism before a pooling layer. A correlation matrix from the co-attention module is supposed to re-assign the weights for dynamically scaling contributions in the magnitude and phase domains according to different pronunciations. Experiments on VoxCeleb showed that the proposed feature-level fusion strategy using the co-attention mechanism gave the Top-1 accuracy of 97.20%, outperforming the state-of-the-art system with 0.82% absolutely, and obtained EER reduction of 0.45% compared to single feature system using FBank.

eess.AS↗

arXiv2025

Investigation of unsupervised and supervised hyperspectral anomaly detection

Hyperspectral sensing is a valuable tool for detecting anomalies and distinguishing between materials in a scene. Hyperspectral anomaly detection (HS-AD) helps characterize the captured scenes and separates them into anomaly and background classes. It is vital in agriculture, environment, and military applications such as RSTA (reconnaissance, surveillance, and target acquisition) missions. We previously designed an equal voting ensemble of hyperspectral unmixing and three unsupervised HS-AD algorithms. We later utilized a supervised classifier to determine the weights of a voting ensemble, creating a hybrid of heterogeneous unsupervised HS-AD algorithms with a supervised classifier in a model stacking, which improved detection accuracy. However, supervised classification methods usually fail to detect novel or unknown patterns that substantially deviate from those seen previously. In this work, we evaluate our technique and other supervised and unsupervised methods using general hyperspectral data to provide new insights.

eess.IV↗

arXiv2025

Misaligned external gas acquisition boosts central black hole activities

One important question in active galactic nucleus (AGN) is how gas is brought down to the galaxy center. Both internal secular evolution (torque induced by non-axisymmetric galactic structures such as bars) and external processes (e.g. mergers or interactions) are expected to redistribute the angular momentum (AM) and transport gas inward. However, it is still under debate whether these processes can significantly affect AGN activities. Here we for the first time report that AGN fraction increases with the difference of kinematic position angles ($ΔPA\equiv|PA_{\mathrm{gas}}-PA_{\mathrm{star}}|$) between ionized gas ($PA_{\mathrm{gas}}$) and stellar disks ($PA_{\mathrm{star}}$) in blue and green galaxies, meanwhile this fraction remains roughly constant for red galaxies. Also the high luminosity AGN fraction increases with $ΔPA$ while the low luminosity AGN fraction is independent with $ΔPA$. These observational results support a scenario in which the interaction between accreted and pre-existing gas provides the AM loss mechanism, thereby the gas inflow fuels the central BH activities, and the AM loss efficiency is positively correlated with the $ΔPA$.

astro-ph.GA↗

arXiv2025

Interface-controlled antiferromagnetic tunnel junctions

Magnetic tunnel junctions (MTJs) are the key building blocks of high-performance spintronic devices. While conventional MTJs rely on ferromagnetic (FM) materials, employing antiferromagnetic (AFM) compounds can significantly increase operation speed and packing density. Current prototypes of AFM tunnel junctions (AFMTJs) exploit antiferromagnets either as spin-filter insulating barriers or as metal electrodes supporting bulk spin-dependent currents. Here, we highlight a largely overlooked AFMTJ prototype, where bulk-spin-degenerate electrodes with an A-type AFM stacking form magnetically uncompensated interfaces, enabling spin-polarized tunneling currents and a sizable tunneling magnetoresistance (TMR) effect. Using first-principles quantum-transport calculations and the van der Waals (vdW) metal Fe$_{4}$GeTe$_{2}$ as a representative A-type AFM electrode, we demonstrate a large negative TMR arising solely from the alignment of interfacial magnetic moments. This prototype of AFMTJs can also be realized with various non-vdW A-type AFM metals that support roughness-insensitive surface magnetization. Beyond TMR, AFMTJs based on A-type antiferromagnets allow convenient switching of the Néel vector, opening a new paradigm for AFM spintronics that leverages spin-dependent properties at AFM interfaces.

cond-mat.mtrl-sci↗

arXiv2025

Ming-Omni: A Unified Multimodal Model for Perception and Generation

We propose Ming-Omni, a unified multimodal model capable of processing images, text, audio, and video, while demonstrating strong proficiency in both speech and image generation. Ming-Omni employs dedicated encoders to extract tokens from different modalities, which are then processed by Ling, an MoE architecture equipped with newly proposed modality-specific routers. This design enables a single model to efficiently process and fuse multimodal inputs within a unified framework, thereby facilitating diverse tasks without requiring separate models, task-specific fine-tuning, or structural redesign. Importantly, Ming-Omni extends beyond conventional multimodal models by supporting audio and image generation. This is achieved through the integration of an advanced audio decoder for natural-sounding speech and Ming-Lite-Uni for high-quality image generation, which also allow the model to engage in context-aware chatting, perform text-to-speech conversion, and conduct versatile image editing. Our experimental results showcase Ming-Omni offers a powerful solution for unified perception and generation across all modalities. Notably, our proposed Ming-Omni is the first open-source model we are aware of to match GPT-4o in modality support, and we release all code and model weights to encourage further research and development in the community.

cs.AI↗

arXiv2025

Multimodal Transformers are Hierarchical Modal-wise Heterogeneous Graphs

Multimodal Sentiment Analysis (MSA) is a rapidly developing field that integrates multimodal information to recognize sentiments, and existing models have made significant progress in this area. The central challenge in MSA is multimodal fusion, which is predominantly addressed by Multimodal Transformers (MulTs). Although act as the paradigm, MulTs suffer from efficiency concerns. In this work, from the perspective of efficiency optimization, we propose and prove that MulTs are hierarchical modal-wise heterogeneous graphs (HMHGs), and we introduce the graph-structured representation pattern of MulTs. Based on this pattern, we propose an Interlaced Mask (IM) mechanism to design the Graph-Structured and Interlaced-Masked Multimodal Transformer (GsiT). It is formally equivalent to MulTs which achieves an efficient weight-sharing mechanism without information disorder through IM, enabling All-Modal-In-One fusion with only 1/3 of the parameters of pure MulTs. A Triton kernel called Decomposition is implemented to ensure avoiding additional computational overhead. Moreover, it achieves significantly higher performance than traditional MulTs. To further validate the effectiveness of GsiT itself and the HMHG concept, we integrate them into multiple state-of-the-art models and demonstrate notable performance improvements and parameter reduction on widely used MSA datasets.

cs.CL↗

arXiv2025

Dynamic QoS Prediction via a Non-Negative Tensor Snowflake Factorization

Dynamic quality of service (QoS) data exhibit rich temporal patterns in user-service interactions, which are crucial for a comprehensive understanding of user behavior and service conditions in Web service. As the number of users and services increases, there is a large amount of unobserved QoS data, which significantly affects users'choice of services. To predict unobserved QoS data, we propose a Non-negative Snowflake Factorization of tensors model. This method designs a snowflake core tensor to enhance the model's learning capability. Additionally, it employs a single latent factor-based, nonnegative multiplication update on tensor (SLF-NMUT) for parameter learning. Empirical results demonstrate that the proposed model more accurately learns dynamic user-service interaction patterns, thereby yielding improved predictions for missing QoS data.

cs.LG↗