arXiv subjects

Nan Yang

Publications and source records attributed to Nan Yang.

At least 73 records · Page 4Linked to original sources

arXiv2025

Magneto-photoelectrochemical 2D heterojunction platform for biosensing detection

Photoelectrochemical (PEC) biosensors exhibit significant potential for biomolecule detection due to their high sensitivity and low background noise. However, their performance is severely constrained by the rapid recombination of photogenerated charge carriers. This study innovatively introduces a non-contact magnetic modulation strategy to suppress electron-hole recombination by manipulating carrier spin states, thereby significantly enhancing photoelectric conversion efficiency. Building on this mechanism, we developed a novel magnetically modulated PEC biosensing platform based on the MXenes/cobalt-doped titanium dioxide (Co-TiO2) heterostructure. This platform achieved ultrasensitive detection of protein kinase A (PKA) activity. Compared to an identical probe-modified biosensor without magnetic field application, the developed platform demonstrated a 68.75% enhancement in detection sensitivity and achieved an ultralow detection limit for PKA of 0.00016 U/mL. It also exhibited a wide linear range from 0.005 to 80 U/mL. This research not only provides a novel methodology for kinase activity analysis but also pioneers the innovative strategy of magnetic modulation for enhanced PEC sensing. It opens new avenues for developing high-performance biosensing platforms, holding significant promise for early disease diagnosis and drug screening applications.

physics.app-ph↗

arXiv2025

MoCa: Modality-aware Continual Pre-training Makes Better Bidirectional Multimodal Embeddings

Multimodal embedding models, built upon causal Vision Language Models (VLMs), have shown promise in various tasks. However, current approaches face three key limitations: the use of causal attention in VLM backbones is suboptimal for embedding tasks; scalability issues due to reliance on high-quality labeled paired data for contrastive learning; and limited diversity in training objectives and data. To address these issues, we propose MoCa, a two-stage framework for transforming pre-trained VLMs into effective bidirectional multimodal embedding models. The first stage, Modality-aware Continual Pre-training, introduces a joint reconstruction objective that simultaneously denoises interleaved text and image inputs, enhancing bidirectional context-aware reasoning. The second stage, Heterogeneous Contrastive Fine-tuning, leverages diverse, semantically rich multimodal data beyond simple image-caption pairs to enhance generalization and alignment. Our method addresses the stated limitations by introducing bidirectional attention through continual pre-training, scaling effectively with massive unlabeled datasets via joint reconstruction objectives, and utilizing diverse multimodal data for enhanced representation robustness. Experiments demonstrate that MoCa consistently improves performance across MMEB and ViDoRe-v2 benchmarks, achieving new state-of-the-art results, and exhibits strong scalability with both model size and training data on MMEB.

cs.CV↗

arXiv2025

4Seasons: A Cross-Season Dataset for Multi-Weather SLAM in Autonomous Driving

We present a novel dataset covering seasonal and challenging perceptual conditions for autonomous driving. Among others, it enables research on visual odometry, global place recognition, and map-based re-localization tracking. The data was collected in different scenarios and under a wide variety of weather conditions and illuminations, including day and night. This resulted in more than 350 km of recordings in nine different environments ranging from multi-level parking garage over urban (including tunnels) to countryside and highway. We provide globally consistent reference poses with up-to centimeter accuracy obtained from the fusion of direct stereo visual-inertial odometry with RTK-GNSS. The full dataset is available at https://go.vision.in.tum.de/4seasons.

cs.CV↗

arXiv2025

4Seasons: Benchmarking Visual SLAM and Long-Term Localization for Autonomous Driving in Challenging Conditions

In this paper, we present a novel visual SLAM and long-term localization benchmark for autonomous driving in challenging conditions based on the large-scale 4Seasons dataset. The proposed benchmark provides drastic appearance variations caused by seasonal changes and diverse weather and illumination conditions. While significant progress has been made in advancing visual SLAM on small-scale datasets with similar conditions, there is still a lack of unified benchmarks representative of real-world scenarios for autonomous driving. We introduce a new unified benchmark for jointly evaluating visual odometry, global place recognition, and map-based visual localization performance which is crucial to successfully enable autonomous driving in any condition. The data has been collected for more than one year, resulting in more than 300 km of recordings in nine different environments ranging from a multi-level parking garage to urban (including tunnels) to countryside and highway. We provide globally consistent reference poses with up to centimeter-level accuracy obtained from the fusion of direct stereo-inertial odometry with RTK GNSS. We evaluate the performance of several state-of-the-art visual odometry and visual localization baseline approaches on the benchmark and analyze their properties. The experimental results provide new insights into current approaches and show promising potential for future research. Our benchmark and evaluation protocols will be available at https://go.vision.in.tum.de/4seasons.

cs.CV↗

arXiv2025

LangYa: Revolutionizing Cross-Spatiotemporal Ocean Forecasting

Ocean forecasting is crucial for both scientific research and societal benefits. Currently, the most accurate forecasting systems are global ocean forecasting systems (GOFSs), which represent the ocean state variables (OSVs) as discrete grids and solve partial differential equations (PDEs) governing the transitions of oceanic state variables using numerical methods. However, GOFSs processes are computationally expensive and prone to cumulative errors. Recently, large artificial intelligence (AI)-based models significantly boosted forecasting speed and accuracy. Unfortunately, building a large AI ocean forecasting system that can be considered cross-spatiotemporal and air-sea coupled forecasts remains a significant challenge. Here, we introduce LangYa, a cross-spatiotemporal and air-sea coupled ocean forecasting system. Results demonstrate that the time embedding module in LangYa enables a single model to make forecasts with lead times ranging from 1 to 7 days. The air-sea coupled module effectively simulates air-sea interactions. The ocean self-attention module improves network stability and accelerates convergence during training, and the adaptive thermocline loss function improves the accuracy of thermocline forecasting. Compared to existing numerical and AI-based ocean forecasting systems, LangYa uses 27 years of global ocean data from the Global Ocean Reanalysis and Simulation version 12 (GLORYS12) for training and achieves more reliable deterministic forecasting results for OSVs. LangYa forecasting system provides global ocean researchers with access to a powerful software tool for accurate ocean forecasting and opens a new paradigm for ocean science.

physics.ao-ph↗

arXiv2025

Sonata: Self-Supervised Learning of Reliable Point Representations

In this paper, we question whether we have a reliable self-supervised point cloud model that can be used for diverse 3D tasks via simple linear probing, even with limited data and minimal computation. We find that existing 3D self-supervised learning approaches fall short when evaluated on representation quality through linear probing. We hypothesize that this is due to what we term the "geometric shortcut", which causes representations to collapse to low-level spatial features. This challenge is unique to 3D and arises from the sparse nature of point cloud data. We address it through two key strategies: obscuring spatial information and enhancing the reliance on input features, ultimately composing a Sonata of 140k point clouds through self-distillation. Sonata is simple and intuitive, yet its learned representations are strong and reliable: zero-shot visualizations demonstrate semantic grouping, alongside strong spatial reasoning through nearest-neighbor relationships. Sonata demonstrates exceptional parameter and data efficiency, tripling linear probing accuracy (from 21.8% to 72.5%) on ScanNet and nearly doubling performance with only 1% of the data compared to previous approaches. Full fine-tuning further advances SOTA across both 3D indoor and outdoor perception tasks.

cs.CV↗

arXiv2025

Bootstrap Your Own Context Length

We introduce a bootstrapping approach to train long-context language models by exploiting their short-context capabilities only. Our method utilizes a simple agent workflow to synthesize diverse long-context instruction tuning data, thereby eliminating the necessity for manual data collection and annotation. The proposed data synthesis workflow requires only a short-context language model, a text retriever, and a document collection, all of which are readily accessible within the open-source ecosystem. Subsequently, language models are fine-tuned using the synthesized data to extend their context lengths. In this manner, we effectively transfer the short-context capabilities of language models to long-context scenarios through a bootstrapping process. We conduct experiments with the open-source Llama-3 family of models and demonstrate that our method can successfully extend the context length to up to 1M tokens, achieving superior performance across various benchmarks.

cs.CL↗

arXiv2025

Identification of Stochastic Gravitational Wave Backgrounds from Cosmic String Using Machine Learning

Cosmic strings play a crucial role in enhancing our understanding of the fundamental structure and evolution of the universe, unifying our knowledge of cosmology, and potentially unveiling new physical laws and phenomena. The advent and operation of space-based detectors provide an important opportunity for detecting stochastic gravitational wave backgrounds (SGWB) generated by cosmic strings. However, the intricate nature of SGWB poses a formidable challenge in distinguishing its signal from the complex noise by some traditional methods. Therefore, we attempt to identify SGWB based on machine learning. Our findings show that the joint detection of LISA and Taiji significantly outperforms individual detectors, and even in the presence of numerous low signal-to-noise ratio(SNR) signals, the identification accuracy remains exceptionally high with 95%. Although our discussion is based solely on simulated data, the relevant methods can provide data-driven analytical capabilities for future observations of SGWB.

gr-qc↗

arXiv2025

Generative Representational Instruction Tuning

All text-based language problems can be reduced to either generation or embedding. Current models only perform well at one or the other. We introduce generative representational instruction tuning (GRIT) whereby a large language model is trained to handle both generative and embedding tasks by distinguishing between them through instructions. Compared to other open models, our resulting GritLM 7B sets a new state of the art on the Massive Text Embedding Benchmark (MTEB) and outperforms all models up to its size on a range of generative tasks. By scaling up further, GritLM 8x7B outperforms all open generative language models that we tried while still being among the best embedding models. Notably, we find that GRIT matches training on only generative or embedding data, thus we can unify both at no performance loss. Among other benefits, the unification via GRIT speeds up Retrieval-Augmented Generation (RAG) by > 60% for long documents, by no longer requiring separate retrieval and generation models. Models, code, etc. are freely available at https://github.com/ContextualAI/gritlm.

cs.CL↗

arXiv2025

Narrowing Information Bottleneck Theory for Multimodal Image-Text Representations Interpretability

The task of identifying multimodal image-text representations has garnered increasing attention, particularly with models such as CLIP (Contrastive Language-Image Pretraining), which demonstrate exceptional performance in learning complex associations between images and text. Despite these advancements, ensuring the interpretability of such models is paramount for their safe deployment in real-world applications, such as healthcare. While numerous interpretability methods have been developed for unimodal tasks, these approaches often fail to transfer effectively to multimodal contexts due to inherent differences in the representation structures. Bottleneck methods, well-established in information theory, have been applied to enhance CLIP's interpretability. However, they are often hindered by strong assumptions or intrinsic randomness. To overcome these challenges, we propose the Narrowing Information Bottleneck Theory, a novel framework that fundamentally redefines the traditional bottleneck approach. This theory is specifically designed to satisfy contemporary attribution axioms, providing a more robust and reliable solution for improving the interpretability of multimodal models. In our experiments, compared to state-of-the-art methods, our approach enhances image interpretability by an average of 9%, text interpretability by an average of 58.83%, and accelerates processing speed by 63.95%. Our code is publicly accessible at https://github.com/LMBTough/NIB.

cs.CV↗

arXiv2025

mmE5: Improving Multimodal Multilingual Embeddings via High-quality Synthetic Data

Multimodal embedding models have gained significant attention for their ability to map data from different modalities, such as text and images, into a unified representation space. However, the limited labeled multimodal data often hinders embedding performance. Recent approaches have leveraged data synthesis to address this problem, yet the quality of synthetic data remains a critical bottleneck. In this work, we identify three criteria for high-quality synthetic multimodal data. First, broad scope ensures that the generated data covers diverse tasks and modalities, making it applicable to various downstream scenarios. Second, robust cross-modal alignment makes different modalities semantically consistent. Third, high fidelity ensures that the synthetic data maintains realistic details to enhance its reliability. Guided by these principles, we synthesize datasets that: (1) cover a wide range of tasks, modality combinations, and languages, (2) are generated via a deep thinking process within a single pass of a multimodal large language model, and (3) incorporate real-world images with accurate and relevant texts, ensuring fidelity through self-evaluation and refinement. Leveraging these high-quality synthetic and labeled datasets, we train a multimodal multilingual E5 model mmE5. Extensive experiments demonstrate that mmE5 achieves state-of-the-art performance on the MMEB Benchmark and superior multilingual performance on the XTD benchmark. Our codes, datasets and models are released in https://github.com/haon-chen/mmE5.

cs.CV↗

arXiv2025

Chaos in Optomechanical Systems coupled to a Non-Markovian environment

We study the chaotic motion of a semi-classical optomechanical system coupled to a non-Markovian environment with a finite correlation time. We show that the non-Markovian environment can significantly enhance chaos, by studying the emergence of chaos using Lyapunov exponent with the changing non-Markovian parameter. It is observed that non-Markovian environment characterized by the Ornstein-Uhlenbeck type noise can modify the generation of chaos with different environmental memory times. As a comparison, the crossover properties from Markov to non-Markovian regimes are also discussed. Our findings indicate that the quantum memory effects on the onset of chaos may become a useful property to be investigated in quantum manipulations and control.

quant-ph↗

arXiv2025

SMamba: Sparse Mamba for Event-based Object Detection

Transformer-based methods have achieved remarkable performance in event-based object detection, owing to the global modeling ability. However, they neglect the influence of non-event and noisy regions and process them uniformly, leading to high computational overhead. To mitigate computation cost, some researchers propose window attention based sparsification strategies to discard unimportant regions, which sacrifices the global modeling ability and results in suboptimal performance. To achieve better trade-off between accuracy and efficiency, we propose Sparse Mamba (SMamba), which performs adaptive sparsification to reduce computational effort while maintaining global modeling capability. Specifically, a Spatio-Temporal Continuity Assessment module is proposed to measure the information content of tokens and discard uninformative ones by leveraging the spatiotemporal distribution differences between activity and noise events. Based on the assessment results, an Information-Prioritized Local Scan strategy is designed to shorten the scan distance between high-information tokens, facilitating interactions among them in the spatial dimension. Furthermore, to extend the global interaction from 2D space to 3D representations, a Global Channel Interaction module is proposed to aggregate channel information from a global spatial perspective. Results on three datasets (Gen1, 1Mpx, and eTram) demonstrate that our model outperforms other methods in both performance and efficiency.

cs.CV↗

arXiv2024

On the Time-Frequency Localization Characteristics of the Delay-Doppler Plane Orthogonal Pulse

In this work, we study the time-frequency (TF) localization characteristics of the prototype pulse of orthogonal delay-Doppler (DD) division multiplexing modulation, namely, the DD plane orthogonal pulse (DDOP). The TF localization characteristics examine how concentrated or spread out the energy of a pulse is in the joint TF domain, the time domain (TD), and the frequency domain (FD). We first derive the TF localization metrics of the DDOP, including its TF area, its time and frequency dispersions, and its direction parameter. Based on these results, we demonstrate that the DDOP exhibits a high energy spread in the TD, FD, and the joint TF domain, while adhering to the Heisenberg uncertainty principle. Thereafter, we discuss the potential advantages brought by the energy spread of the DDOP, especially with regard to harnessing both time and frequency diversities and enabling fine-resolution sensing. Subsequently, we examine the relationships between the time and frequency dispersions of the DDOP and those of the envelope functions of DDOP's TD and FD representations, paving the way for simplified determination of the TF localization metrics for more generalized variants of the DDOP and the pulses used in other DD domain modulation schemes. Finally, using numerical results, we validate our analysis and find further insights.

eess.SP↗

arXiv2024

LongEmbed: Extending Embedding Models for Long Context Retrieval

Embedding models play a pivot role in modern NLP applications such as IR and RAG. While the context limit of LLMs has been pushed beyond 1 million tokens, embedding models are still confined to a narrow context window not exceeding 8k tokens, refrained from application scenarios requiring long inputs such as legal contracts. This paper explores context window extension of existing embedding models, pushing the limit to 32k without requiring additional training. First, we examine the performance of current embedding models for long context retrieval on our newly constructed LongEmbed benchmark. LongEmbed comprises two synthetic tasks and four carefully chosen real-world tasks, featuring documents of varying length and dispersed target information. Benchmarking results underscore huge room for improvement in these models. Based on this, comprehensive experiments show that training-free context window extension strategies like position interpolation can effectively extend the context window of existing embedding models by several folds, regardless of their original context being 512 or beyond 4k. Furthermore, for models employing absolute position encoding (APE), we show the possibility of further fine-tuning to harvest notable performance gains while strictly preserving original behavior for short inputs. For models using rotary position embedding (RoPE), significant enhancements are observed when employing RoPE-specific methods, such as NTK and SelfExtend, indicating RoPE's superiority over APE for context window extension. To facilitate future research, we release E5-Base-4k and E5-RoPE-Base, along with the LongEmbed benchmark.

cs.CL↗

arXiv2024

Little Giants: Synthesizing High-Quality Embedding Data at Scale

Synthetic data generation has become an increasingly popular way of training models without the need for large, manually labeled datasets. For tasks like text embedding, synthetic data offers diverse and scalable training examples, significantly reducing the cost of human annotation. However, most current approaches rely heavily on proprietary models like GPT-4, which are expensive and inefficient for generating large-scale embedding data. In this paper, we introduce SPEED, a framework that aligns open-source small models (8B) to efficiently generate large-scale synthetic embedding data. Through supervised fine-tuning, preference optimization, and self-improvement, SPEED enables small open-source models to produce high-quality data. Remarkably, SPEED uses only less than 1/10 of the GPT API calls, outperforming the state-of-the-art embedding model E5_mistral when both are trained solely on their synthetic data. Using this efficient generator, we conduct a comprehensive study on how various factors within the alignment pipeline impact data quality and reveal the scaling law for synthetic embedding data.

cs.CL↗

arXiv2024

CAKD: A Correlation-Aware Knowledge Distillation Framework Based on Decoupling Kullback-Leibler Divergence

In knowledge distillation, a primary focus has been on transforming and balancing multiple distillation components. In this work, we emphasize the importance of thoroughly examining each distillation component, as we observe that not all elements are equally crucial. From this perspective,we decouple the Kullback-Leibler (KL) divergence into three unique elements: Binary Classification Divergence (BCD), Strong Correlation Divergence (SCD), and Weak Correlation Divergence (WCD). Each of these elements presents varying degrees of influence. Leveraging these insights, we present the Correlation-Aware Knowledge Distillation (CAKD) framework. CAKD is designed to prioritize the facets of the distillation components that have the most substantial influence on predictions, thereby optimizing knowledge transfer from teacher to student models. Our experiments demonstrate that adjusting the effect of each element enhances the effectiveness of knowledge transformation. Furthermore, evidence shows that our novel CAKD framework consistently outperforms the baseline across diverse models and datasets. Our work further highlights the importance and effectiveness of closely examining the impact of different parts of distillation process.

cs.LG↗

arXiv2024

Coverage Analysis for 3D Indoor Terahertz Communication System Over Fluctuating Two-Ray Fading Channels

In this paper, we develop a novel analytical framework for a three-dimensional (3D) indoor terahertz (THz) communication system. Our proposed model incorporates more accurate modeling of wall blockages via Manhattan line processes and precise modeling of THz fading channels via a fluctuating two-ray (FTR) channel model. We also account for traditional unique features of THz, such as molecular absorption loss, user blockages, and 3D directional antenna beams. Moreover, we model locations of access points (APs) using a Poisson point process and adopt the nearest line-of-sight AP association strategy. Due to the high penetration loss caused by wall blockages, we consider that a user equipment (UE) and its associated AP and interfering APs are all in the same rectangular area, i.e., a room. Based on the proposed rectangular area model, we evaluate the impact of the UE's location on the distance to its associated AP. We then develop a tractable method to derive a new expression for the coverage probability by examining the interference from interfering APs and considering the FTR fading experienced by THz communications. Aided by simulation results, we validate our analysis and demonstrate that the UE's location has a pronounced impact on its coverage probability. Additionally, we find that the optimal AP density is determined by both the UE's location and the room size, which provides valuable insights for meeting the coverage requirements of future THz communication system deployment.

cs.IT↗