Searcharxiv⌕ Search

arXiv subjects

Zehao Lin

Publications and source records attributed to Zehao Lin.

At least 37 records · Page 2Linked to original sources

TAdaRAG: Task Adaptive Retrieval-Augmented Generation via On-the-Fly Knowledge Graph Construction

Retrieval-Augmented Generation (RAG) improves large language models by retrieving external knowledge, often truncated into smaller chunks due to the input context window, which leads to information loss, resulting in response hallucinations and broken reasoning chains. Moreover, traditional RAG retrieves unstructured knowledge, introducing irrelevant details that hinder accurate reasoning. To address these issues, we propose TAdaRAG, a novel RAG framework for on-the-fly task-adaptive knowledge graph construction from external sources. Specifically, we design an intent-driven routing mechanism to a domain-specific extraction template, followed by supervised fine-tuning and a reinforcement learning-based implicit extraction mechanism, ensuring concise, coherent, and non-redundant knowledge integration. Evaluations on six public benchmarks and a real-world business benchmark (NowNewsQA) across three backbone models demonstrate that TAdaRAG outperforms existing methods across diverse domains and long-text tasks, highlighting its strong generalization and practical effectiveness.

cs.CL↗

How Does Environmental Information Disclosure Affect Corporate Environmental Performance? Evidence from Chinese A-Share Listed Companies

Global climate warming and air pollution pose severe threats to economic development and public safety, presenting significant challenges to sustainable development worldwide. Corporations, as key players in resource utilization and emissions, have drawn increasing attention from policymakers, researchers, and the public regarding their environmental strategies and practices. This study employs a two-way fixed effects panel model to examine the impact of environmental information disclosure on corporate environmental performance, its regional heterogeneity, and the underlying mechanisms. The results demonstrate that environmental information disclosure significantly improves corporate environmental performance, with the effect being more pronounced in areas of high population density and limited green space. These findings provide empirical evidence supporting the role of environmental information disclosure as a critical tool for improving corporate environmental practices. The study highlights the importance of targeted, region-specific policies to maximize the effectiveness of disclosure, offering valuable insights for promoting sustainable development through enhanced corporate transparency.

econ.GN↗

Innovator: Scientific Continued Pretraining with Fine-grained MoE Upcycling

A large language model (LLM) with knowledge in both scientific and general tasks is the foundation of science general intelligence. However, directly continued pretraining an LLM using science data usually leads to catastrophic forgetting, which indicates severe degradation in general ability. In this report, we present Innovator, which solves this problem by upcycling a pre-trained dense LLM into a fine-grained Mixtures-of-Experts model during continued pretraining, where different experts are expected to learn science knowledge in different disciplines, and a shared expert is utilized for general tasks. Innovator introduces a four-stage upcycle training paradigm: (1) Scientific Expert Induction on discipline-specific data, (2) Fine-grained Expert Splitting via FFN dimension decomposition, (3) Science-Aware Routing warmup, and (4) Generalist-Scientist Integration training on hybrid datasets. Such a paradigm enables knowledge in the general domain, and different scientific disciplines can be decoupled, avoiding the negative influence among knowledge in different domains. With 53.3B total parameters and 13.3B activated, Innovator extends Qwen2.5-7B using a shared general expert and 64 specialized scientific experts with 8 activated. Trained on 300B tokens with tri-level quality-controlled data, Innovator achieves 25% average improvement across 30 scientific tasks with a win rate as 70%, while retaining 99% performance in general tasks. Furthermore, Innovator-Reason, which is post-trained from Innovator for reasoning boosting, exhibits excellent reasoning performance in solving complex scientific problems with improvements over 30%.

cs.LG↗

TA-VLA: Elucidating the Design Space of Torque-aware Vision-Language-Action Models

Many robotic manipulation tasks require sensing and responding to force signals such as torque to assess whether the task has been successfully completed and to enable closed-loop control. However, current Vision-Language-Action (VLA) models lack the ability to integrate such subtle physical feedback. In this work, we explore Torque-aware VLA models, aiming to bridge this gap by systematically studying the design space for incorporating torque signals into existing VLA architectures. We identify and evaluate several strategies, leading to three key findings. First, introducing torque adapters into the decoder consistently outperforms inserting them into the encoder.Third, inspired by joint prediction and planning paradigms in autonomous driving, we propose predicting torque as an auxiliary output, which further improves performance. This strategy encourages the model to build a physically grounded internal representation of interaction dynamics. Extensive quantitative and qualitative experiments across contact-rich manipulation benchmarks validate our findings.

cs.RO↗

Scalable Complexity Control Facilitates Reasoning Ability of LLMs

The reasoning ability of large language models (LLMs) has been rapidly advancing in recent years, attracting interest in more fundamental approaches that can reliably enhance their generalizability. This work demonstrates that model complexity control, conveniently implementable by adjusting the initialization rate and weight decay coefficient, improves the scaling law of LLMs consistently over varying model sizes and data sizes. This gain is further illustrated by comparing the benchmark performance of 2.4B models pretrained on 1T tokens with different complexity hyperparameters. Instead of fixing the initialization std, we found that a constant initialization rate (the exponent of std) enables the scaling law to descend faster in both model and data sizes. These results indicate that complexity control is a promising direction for the continual advancement of LLMs.

cs.LG↗

Xinyu AI Search: Enhanced Relevance and Comprehensive Results with Rich Answer Presentations

Traditional search engines struggle to synthesize fragmented information for complex queries, while generative AI search engines face challenges in relevance, comprehensiveness, and presentation. To address these limitations, we introduce Xinyu AI Search, a novel system that incorporates a query-decomposition graph to dynamically break down complex queries into sub-queries, enabling stepwise retrieval and generation. Our retrieval pipeline enhances diversity through multi-source aggregation and query expansion, while filtering and re-ranking strategies optimize passage relevance. Additionally, Xinyu AI Search introduces a novel approach for fine-grained, precise built-in citation and innovates in result presentation by integrating timeline visualization and textual-visual choreography. Evaluated on recent real-world queries, Xinyu AI Search outperforms eight existing technologies in human assessments, excelling in relevance, comprehensiveness, and insightfulness. Ablation studies validate the necessity of its key sub-modules. Our work presents the first comprehensive framework for generative AI search engines, bridging retrieval, generation, and user-centric presentation.

cs.IR↗

MemOS: An Operating System for Memory-Augmented Generation (MAG) in Large Language Models

Large Language Models (LLMs) have emerged as foundational infrastructure in the pursuit of Artificial General Intelligence (AGI). Despite their remarkable capabilities in language perception and generation, current LLMs fundamentally lack a unified and structured architecture for handling memory. They primarily rely on parametric memory (knowledge encoded in model weights) and ephemeral activation memory (context-limited runtime states). While emerging methods like Retrieval-Augmented Generation (RAG) incorporate plaintext memory, they lack lifecycle management and multi-modal integration, limiting their capacity for long-term knowledge evolution. To address this, we introduce MemOS, a memory operating system designed for LLMs that, for the first time, elevates memory to a first-class operational resource. It builds unified mechanisms for representation, organization, and governance across three core memory types: parametric, activation, and plaintext. At its core is the MemCube, a standardized memory abstraction that enables tracking, fusion, and migration of heterogeneous memory, while offering structured, traceable access across tasks and contexts. MemOS establishes a memory-centric execution framework with strong controllability, adaptability, and evolvability. It fills a critical gap in current LLM infrastructure and lays the groundwork for continual adaptation, personalized intelligence, and cross-platform coordination in next-generation intelligent systems.

cs.CL↗

Experimental Investigation of Variations in Polycrystalline Hf0.5Zr0.5O2 (HZO)-based MFIM

Device-to-device variations in ferroelectric (FE) hafnium oxide (HfO2)-based devices pose a crucial challenge that limits the otherwise promising capabilities of this technology. Although previous simulation-based studies have identified polarization (P) domain nucleation and polycrystallinity as key contributors to variations in HfO2, experimental validation remains limited. Here, we experimentally investigate variations in remanent polarization (PR) of Hf0.5Zr0.5O2 (HZO)-based metal-ferroelectric-insulator-metal (MFIM) capacitors across different set voltages (VSET) and FE thicknesses (TFE). Our measurements reveal a non-monotonic behavior of the standard deviation of PR with VSET peaking around coercive voltage (VC), which is consistent with previous simulation-based predictions. In the low- and high-VSET regions, PR variations are primarily dictated by saturation polarization (PS) variations, mainly originating from charge trap effects at the interface between the FE-dielectric (DE) layer and the polycrystallinity of FE. On the other hand, in the mid-VSET region peak, the PR variations are attributed to the VC variation, which comes from a combined effect of multi-domain (MD) P switching and polycrystallinity. Notably, sharp P switching associated with domain nucleation amplifies the variations, resulting in a peak of PR variations in this VSET range. Further, we observe that as HZO thickness (TFE) is scaled, the non-monotonicity in variations with VSET is reduced, primarily due to reduced domain nucleation and smaller grain sizes. We experimentally establish a strong correlation of PR with PS in the low- and high-VSET regions and with VC in the mid-VSET region across various TFE. Finally, our experimental findings are corroborated with simulations using a 3D phase-field model.

physics.app-ph↗

Calibrating the Color-Magnitude Relation of M Giants by Using Open Clusters

M giants, with their distinctive properties such as high luminosity, serve as excellent indicators for mapping the structure of the Milky Way. The distance to distant M giants can be determined by using the color-magnitude relation (CMR), which is derived from color-magnitude diagrams of specific systems in previous studies. In this work, we aimed to achieve more accurate distance determination for M giants by focusing on open clusters (OCs) with a large number of member stars and thus improve the CMR. For the first time, we compiled a census of OCs harboring M giants using Gaia Data Release 3 (DR3) and Large Sky Area Multi-Object Fiber Spectroscopic Telescope Data Release 9. We identified 58 M giants associated with 43 OCs and obtained their astrometric and photometric parameters from Gaia DR3. Using the distances of these OCs, we derived the CMR for M giants as a linear correlation, expressed as $M_{Ks}=3.85-8.26(J-K_s$). This linear relation proved superior to the empirical distance relation in characterizing the CMR of M giants. The photometric distances of M giants derived from the CMR are consistent with the parallax distances from Gaia and known spectroscopic distances, with median deviations of 1.5% and 2.3%, respectively. Using the distances of M giants derived from the CMR, we computed their radial velocity ($V_R$), azimuthal velocity ($Vϕ$), and vertical velocity ($V_Z$), respectively. The distributions of these velocities revealed key features of the Galactic disk, including oscillation, north-south rotational asymmetry, and warp. These findings are consistent with previous studies and further validate the reliability of the derived CMR.

astro-ph.SR↗

SurveyX: Academic Survey Automation via Large Language Models

Large Language Models (LLMs) have demonstrated exceptional comprehension capabilities and a vast knowledge base, suggesting that LLMs can serve as efficient tools for automated survey generation. However, recent research related to automated survey generation remains constrained by some critical limitations like finite context window, lack of in-depth content discussion, and absence of systematic evaluation frameworks. Inspired by human writing processes, we propose SurveyX, an efficient and organized system for automated survey generation that decomposes the survey composing process into two phases: the Preparation and Generation phases. By innovatively introducing online reference retrieval, a pre-processing method called AttributeTree, and a re-polishing process, SurveyX significantly enhances the efficacy of survey composition. Experimental evaluation results show that SurveyX outperforms existing automated survey generation systems in content quality (0.259 improvement) and citation quality (1.76 enhancement), approaching human expert performance across multiple evaluation dimensions. Examples of surveys generated by SurveyX are available on www.surveyx.cn

cs.CL↗

Ultralow Voltage Operation of p- and n-FETs Enabled by Self-Formed Gate Dielectric and Metal Contacts on 2D Tellurium

The ongoing demand for more energy-efficient, high-performance electronics is driving the exploration of innovative materials and device architectures, where interfaces play a crucial role due to the continuous downscaling of device dimensions. Tellurium (Te), in its two-dimensional (2D) form, offers significant potential due to its high carrier mobility and ambipolar characteristics, with the carrier type easily tunable via surface modulation. In this study, we leverage atomically controlled material transformations in 2D Te to create intimate junctions, enabling near-ideal field-effect transistors (FETs) for both n-type and p-type operation. A NiTex-Te contact provides highly transparent interfaces, resulting in low contact resistance, while the TiOx-Te gate dielectric forms an ultraclean interface with a capacitance equivalent to 0.88 nm equivalent oxide thickness (EOT), where the quantum capacitance of Te is observed. Subthreshold slopes (SS) approach the Boltzmann limit, with a record-low SS of 3.5 mV/dec achieved at 10 K. Furthermore, we demonstrate 2D Te-based complementary metal-oxide-semiconductor (CMOS) inverters operating at an ultralow voltage of 0.08 V with a voltage gain of 7.1 V/V. This work presents a promising approach to forming intimate dielectric/semiconductor and metal/semiconductor junctions for next-generation low-power electronic devices.

physics.app-ph↗

FTII-Bench: A Comprehensive Multimodal Benchmark for Flow Text with Image Insertion

Benefiting from the revolutionary advances in large language models (LLMs) and foundational vision models, large vision-language models (LVLMs) have also made significant progress. However, current benchmarks focus on tasks that evaluating only a single aspect of LVLM capabilities (e.g., recognition, detection, understanding). These tasks fail to fully demonstrate LVLMs' potential in complex application scenarios. To comprehensively assess the performance of existing LVLMs, we propose a more challenging task called the Flow Text with Image Insertion task (FTII). This task requires LVLMs to simultaneously possess outstanding abilities in image comprehension, instruction understanding, and long-text interpretation. Specifically, given several text paragraphs and a set of candidate images, as the text paragraphs accumulate, the LVLMs are required to select the most suitable image from the candidates to insert after the corresponding paragraph. Constructing a benchmark for such a task is highly challenging, particularly in determining the sequence of flowing text and images. To address this challenge, we turn to professional news reports, which naturally contain a gold standard for image-text sequences. Based on this, we introduce the Flow Text with Image Insertion Benchmark (FTII-Bench), which includes 318 high-quality Chinese image-text news articles and 307 high-quality English image-text news articles, covering 10 different news domains. Using these 625 high-quality articles, we construct problems of two different types with multiple levels of difficulty. Furthermore, we establish two different evaluation pipelines based on the CLIP model and existing LVLMs. We evaluate 9 open-source and 2 closed-source LVLMs as well as 2 CLIP-based models. Results indicate that even the most advanced models (e.g., GPT-4o) face significant challenges when tackling the FTII task.

cs.CV↗

Kinematics of Young Stellar Objects Under Various Stellar Feedback

Based on the Gaia Data Release 3 and APOGEE datasets, we investigate the kinematic differences between young stellar objects (YSOs) and their parent clouds in five nearby star-forming regions. Overall, the 1D velocity differences between Class II YSOs and their parent molecular cloud range from [0, 1.4] km/s. In feedback environments dominated by outflows, massive stars, and supernova feedback, the corresponding velocity differences range from [0, 1.4] km/s, [0.1, 0.4] km/s, and [0.1, 1] km/s, respectively. These results indicate that YSO kinematics are not significantly affected by these different types of feedback environment. Additionally, compared to the Class II YSOs, Class III YSOs have slightly larger velocities and dispersions.

astro-ph.GA↗

A New Statistical Analysis of the Morphology of Spiral Galaxies

Morphology is the starting point for understanding galaxies. Elmegreen et al. classified spiral galaxies into flocculent, multiple-arm, and grand-design galaxies based on the regularity of their spiral arm structure. With the release of a vast number of clear spiral galaxy images from the Sloan Digital Sky Survey, we conducted a morphological classification of 5093 blue spiral galaxies. A statistical analysis of this sample shows that the fractions of flocculent, multiple-arm, and grand-design galaxies are 38 $\pm$ 1%, 59 $\pm$ 1%, and 3 $\pm$ 1%, respectively. Redshift has no obvious influence on this classification. However, as the bulge size becomes larger, the fraction of multiple-arm galaxies increases, while that of flocculent galaxies decreases. In addition, we performed a statistical analysis of 3958 galaxies with a clear spiral arm structure, finding 82% of these galaxies have two arms in their inner regions. We also found that the majority (74%) of the barred spiral galaxies exhibit the characteristics of two inner spiral arms and multiple outer spiral arms, and there is no barred spiral galaxy in this work with four continuous spiral arms from the inner to the outer regions. These results highlight that the spiral arm structure of the Milky Way, according to the current mainstream view of a four-arm galaxy with continuous arms extending from the inner to outer regions, is quite unique. However, our findings align with the spiral morphology of the Milky Way proposed by Xu et al., in which case our Galaxy can be considered typical.

astro-ph.GA↗

MM-CamObj: A Comprehensive Multimodal Dataset for Camouflaged Object Scenarios

Large visual-language models (LVLMs) have achieved great success in multiple applications. However, they still encounter challenges in complex scenes, especially those involving camouflaged objects. This is primarily due to the lack of samples related to camouflaged scenes in the training dataset. To mitigate this issue, we construct the MM-CamObj dataset for the first time, comprising two subsets: CamObj-Align and CamObj-Instruct. Specifically, CamObj-Align contains 11,363 image-text pairs, and it is designed for VL alignment and injecting rich knowledge of camouflaged scenes into LVLMs. CamObj-Instruct is collected for fine-tuning the LVLMs with improved instruction-following capabilities, and it includes 11,363 images and 68,849 conversations with diverse instructions. Based on the MM-CamObj dataset, we propose the CamObj-Llava, an LVLM specifically designed for addressing tasks in camouflaged scenes. To facilitate our model's effective acquisition of knowledge about camouflaged objects and scenes, we introduce a curriculum learning strategy with six distinct modes. Additionally, we construct the CamObj-Bench to evaluate the existing LVLMs' capabilities of understanding, recognition, localization and count in camouflage scenes. This benchmark includes 600 images and 7 tasks, with a total of 9,449 questions. Extensive experiments are conducted on the CamObj-Bench with CamObj-Llava, 8 existing open-source and 3 closed-source LVLMs. Surprisingly, the results indicate that our model achieves a 25.84% improvement in 4 out of 7 tasks compared to GPT-4o. Code and datasets will be available at https://github.com/JCruan519/MM-CamObj.

cs.CV↗

Search for Classical Cepheids in Galactic Open Clusters and Calibration of the Period Wesenheit Metallicity Relation in the Gaia Bands

It is beneficial to calibrate the period Wesenheit metallicity relation (PWZR) of Delta Cephei stars (DCEPs), i.e., classical Cepheids, using accurate parallaxes of associated open clusters (OCs) from Gaia data release 3 (DR3). To this aim, we obtain a total of 43 OC-DCEPs (including 33 fundamental mode, 9 first overtone mode, and 1 multimode DCEPs.) and calibrate the PWZR as $W_G=(-3.356 \,\pm\, 0.033) \,(\log{P-1})+(-5.947 \,\pm\, 0.025)+(-0.285 \,\pm\, 0.064)[\textrm{Fe/H}]$. The concurrently obtained residual parallax offset in OC, $zp = -4\pm5\,μ\textrm{as}$, demonstrate the adequacy of the parallax corrections within the magnitude range of OC member stars. By comparing the field DCEPs' DR3 parallaxes with their photometric parallaxes derived by our PWZR, we estimated the residual parallax offset in field DCEPs as $zp = -15\pm3\,μ\textrm{as}$. Using our PWZR, we estimate the distance modulus of the Large Magellanic Cloud to be $18.482 \,\pm\, 0.040$ mag, which aligns well with the most accurate published value obtained through geometric methods.

astro-ph.SR↗

$\text{Memory}^3$: Language Modeling with Explicit Memory

The training and inference of large language models (LLMs) are together a costly process that transports knowledge from raw data to meaningful computation. Inspired by the memory hierarchy of the human brain, we reduce this cost by equipping LLMs with explicit memory, a memory format cheaper than model parameters and text retrieval-augmented generation (RAG). Conceptually, with most of its knowledge externalized to explicit memories, the LLM can enjoy a smaller parameter size, training cost, and inference cost, all proportional to the amount of remaining "abstract knowledge". As a preliminary proof of concept, we train from scratch a 2.4B LLM, which achieves better performance than much larger LLMs as well as RAG models, and maintains higher decoding speed than RAG. The model is named $\text{Memory}^3$, since explicit memory is the third form of memory in LLMs after implicit memory (model parameters) and working memory (context key-values). We introduce a memory circuitry theory to support the externalization of knowledge, and present novel techniques including a memory sparsification mechanism that makes storage tractable and a two-stage pretraining scheme that facilitates memory formation.

cs.CL↗

VLBI with SKA: Possible Arrays and Astrometric Science

The next generation of very long baseline interferometry (VLBI) is stepping into the era of microarcsecond ($μ$as) astronomy, and pushing astronomy, especially astrometry, to new heights. VLBI with the Square Kilometre Array (SKA), SKA-VLBI, will increase current sensitivity by an order of magnitude, and reach astrometric precision routinely below 10 $μ$as, even challenging 1 $μ$as. This advancement allows precise parallax and proper motion measurements of various celestial objects. Such improvements can be used to study objects (including isolated objects, and binary or multiple systems) in different stellar stages (such as star formation, main-sequence stars, asymptotic giant branch stars, pulsars, black holes, white dwarfs, etc.), unveil the structure and evolution of complex systems (such as the Milky Way), benchmark the international celestial reference frame, and reveal cosmic expansion. Furthermore, the theory of general relativity can also be tested with SKA-VLBI using precise measurements of light deflection under the gravitational fields of different solar system objects and the perihelion precession of solar system objects.

astro-ph.GA↗