arXiv subjects

Song Jin

Publications and source records attributed to Song Jin.

At least 19 recordsLinked to original sources

arXiv2026

Allocate Before You Embed: Adaptive Visual Input Allocation for Video Embeddings

Large-scale video retrieval requires embedding models to encode long and diverse videos under tight visual-input and inference budgets. Existing methods typically sample a small, fixed set of frames at their original resolution, limiting temporal coverage and ignoring frame importance. Our empirical analysis shows that expanding temporal coverage improves retrieval even under a fixed visual-input budget. Gains are larger when the original per-frame resolution is preserved, highlighting the complementary roles of temporal coverage and spatial fidelity. Motivated by this finding, we propose AllocEmbed, an allocate-then-embed framework that reallocates a fixed visual-input budget across more frames. A lightweight allocator uses low-cost previews to assign frame-wise resolutions before the embedding backbone, preserving more detail where it most benefits retrieval while reducing visual cost elsewhere. We further introduce Retrieval-Driven Policy Optimization (RDPO), which learns the allocator directly from retrieval feedback using a rank-validated similarity gap and a confidence-guided efficiency incentive. Operating entirely before the backbone, AllocEmbed integrates with existing retrieval systems without modifying the embedding model or downstream pipeline. Experiments on the MMEB-V2 V-QA and V-RET tasks and our LongRet benchmark show that AllocEmbed achieves the best overall retrieval performance among the evaluated budget-matched methods and transfers across embedding backbones. Our code is publicly available at https://github.com/jinsong8/AllocEmbed.

cs.CV↗

arXiv2026

DiningBench: A Hierarchical Multi-view Benchmark for Perception and Reasoning in the Dietary Domain

Recent advancements in Vision-Language Models (VLMs) have revolutionized general visual understanding. However, their application in the food domain remains constrained by benchmarks that rely on coarse-grained categories, single-view imagery, and inaccurate metadata. To bridge this gap, we introduce DiningBench, a hierarchical, multi-view benchmark designed to evaluate VLMs across three levels of cognitive complexity: Fine-Grained Classification, Nutrition Estimation, and Visual Question Answering. Unlike previous datasets, DiningBench comprises 3,021 distinct dishes with an average of 5.27 images per entry, incorporating fine-grained "hard" negatives from identical menus and rigorous, verification-based nutritional data. We conduct an extensive evaluation of 29 state-of-the-art open-source and proprietary models. Our experiments reveal that while current VLMs excel at general reasoning, they struggle significantly with fine-grained visual discrimination and precise nutritional reasoning. Furthermore, we systematically investigate the impact of multi-view inputs and Chain-of-Thought reasoning, identifying five primary failure modes. DiningBench serves as a challenging testbed to drive the next generation of food-centric VLM research. All codes are released in https://github.com/meituan/DiningBench.

cs.CV↗

arXiv2025

FinRpt: Dataset, Evaluation System and LLM-based Multi-agent Framework for Equity Research Report Generation

While LLMs have shown great success in financial tasks like stock prediction and question answering, their application in fully automating Equity Research Report generation remains uncharted territory. In this paper, we formulate the Equity Research Report (ERR) Generation task for the first time. To address the data scarcity and the evaluation metrics absence, we present an open-source evaluation benchmark for ERR generation - FinRpt. We frame a Dataset Construction Pipeline that integrates 7 financial data types and produces a high-quality ERR dataset automatically, which could be used for model training and evaluation. We also introduce a comprehensive evaluation system including 11 metrics to assess the generated ERRs. Moreover, we propose a multi-agent framework specifically tailored to address this task, named FinRpt-Gen, and train several LLM-based agents on the proposed datasets using Supervised Fine-Tuning and Reinforcement Learning. Experimental results indicate the data quality and metrics effectiveness of the benchmark FinRpt and the strong performance of FinRpt-Gen, showcasing their potential to drive innovation in the ERR generation field. All code and datasets are publicly available.

cs.CL↗

arXiv2025

ViPER: Empowering the Self-Evolution of Visual Perception Abilities in Vision-Language Model

The limited capacity for fine-grained visual perception presents a critical bottleneck for Vision-Language Models (VLMs) in real-world applications. Addressing this is challenging due to the scarcity of high-quality data and the limitations of existing methods: supervised fine-tuning (SFT) often compromises general capabilities, while reinforcement fine-tuning (RFT) prioritizes textual reasoning over visual perception. To bridge this gap, we propose a novel two-stage task that structures visual perception learning as a coarse-to-fine progressive process. Based on this task formulation, we develop ViPER, a self-bootstrapping framework specifically designed to enable iterative evolution through self-critiquing and self-prediction. By synergistically integrating image-level and instance-level reconstruction with a two-stage reinforcement learning strategy, ViPER establishes a closed-loop training paradigm, where internally synthesized data directly fuel the enhancement of perceptual ability. Applied to the Qwen2.5-VL family, ViPER produces the Qwen-Viper series. With an average gain of 1.7% on seven comprehensive benchmarks spanning various tasks and up to 6.0% on fine-grained perception, Qwen-Viper consistently demonstrates superior performance across different vision-language scenarios while maintaining generalizability. Beyond enabling self-improvement in perceptual capabilities, ViPER provides concrete evidence for the reciprocal relationship between generation and understanding, a breakthrough to developing more autonomous and capable VLMs.

cs.CV↗

arXiv2025

Tagging the Thought: Unlocking Personalization Reasoning via Reinforcement Learning

Recent advancements have endowed Large Language Models (LLMs) with impressive general reasoning capabilities, yet they often struggle with personalization reasoning - the crucial ability to analyze user history, infer unique preferences, and generate tailored responses. To address this limitation, we introduce TagPR, a novel training framework that significantly enhances an LLM's intrinsic capacity for personalization reasoning through a tagging the thought approach. Our method first develops a data-driven pipeline to automatically generate and semantically label reasoning chains, creating a structured dataset that fosters interpretable reasoning. We then propose a synergistic training strategy that begins with Supervised Fine-Tuning (SFT) on this tagged data to establish foundational reasoning patterns, followed by a multi-stage reinforcement learning (RL) process. This RL phase is guided by a unique composite reward signal, which integrates tag-based constraints and a novel Personalization Reward Model with User Embeddings (PRMU) to achieve fine-grained alignment with user-specific logic. Extensive experiments on the public LaMP benchmark and a self-constructed dataset demonstrate that our approach achieves state-of-the-art results, delivering an average improvement of 32.65% over the base model across all tasks. Our work validates that structured, interpretable reasoning is a highly effective pathway to unlocking genuine personalization capabilities in LLMs.

cs.CL↗

arXiv2025

Flat band excitons in a three-dimensional supertwisted spiral transition metal dichalcogenide

A new frontier in van der Waals twistronics is the development of three-dimensional (3D) supertwisted materials, where each successive atomic layer rotates by the same angle. While two-dimensional (2D) moire systems have been extensively studied, the unique phenomena arising from 3D twistronics remain largely unexplored. In this work, we report the discovery of flat-band excitons in 3D supertwisted WS2, revealed by systematic photoluminescence (PL) experiments and electronic structure calculations. These excitons retain key features of 2D moire transition metal dichalcogenides (TMDs)-such as layer confinement, moire-driven localization, and strong Coulomb interactions-while also offering advantages in scalability and enhanced optical responses in three dimensions. Beyond the PL signatures reminiscent of 2D A excitons, we observe novel direct and indirect exciton emission uniquely tied to the supertwist geometry. Using generalized Bloch band theory and local density of states calculations that incorporate screw rotational symmetry, we uncovered the coexistence of 2D and 3D flatband gaps. These flat-band excitons serve as sensitive probes of the electronic properties of 3D supertwisted semiconductors and open new pathways for applications in quantum optoelectronics.

physics.app-ph↗

arXiv2025

Beyond Static Testbeds: An Interaction-Centric Agent Simulation Platform for Dynamic Recommender Systems

Evaluating and iterating upon recommender systems is crucial, yet traditional A/B testing is resource-intensive, and offline methods struggle with dynamic user-platform interactions. While agent-based simulation is promising, existing platforms often lack a mechanism for user actions to dynamically reshape the environment. To bridge this gap, we introduce RecInter, a novel agent-based simulation platform for recommender systems featuring a robust interaction mechanism. In RecInter platform, simulated user actions (e.g., likes, reviews, purchases) dynamically update item attributes in real-time, and introduced Merchant Agents can reply, fostering a more realistic and evolving ecosystem. High-fidelity simulation is ensured through Multidimensional User Profiling module, Advanced Agent Architecture, and LLM fine-tuned on Chain-of-Thought (CoT) enriched interaction data. Our platform achieves significantly improved simulation credibility and successfully replicates emergent phenomena like Brand Loyalty and the Matthew Effect. Experiments demonstrate that this interaction mechanism is pivotal for simulating realistic system evolution, establishing our platform as a credible testbed for recommender systems research. Our codes are available at https://github.com/jinsong8/RecInter.

cs.CL↗

arXiv2024

PEER: Expertizing Domain-Specific Tasks with a Multi-Agent Framework and Tuning Methods

In domain-specific applications, GPT-4, augmented with precise prompts or Retrieval-Augmented Generation (RAG), shows notable potential but faces the critical tri-lemma of performance, cost, and data privacy. High performance requires sophisticated processing techniques, yet managing multiple agents within a complex workflow often proves costly and challenging. To address this, we introduce the PEER (Plan, Execute, Express, Review) multi-agent framework. This systematizes domain-specific tasks by integrating precise question decomposition, advanced information retrieval, comprehensive summarization, and rigorous self-assessment. Given the concerns of cost and data privacy, enterprises are shifting from proprietary models like GPT-4 to custom models, striking a balance between cost, security, and performance. We developed industrial practices leveraging online data and user feedback for efficient model tuning. This study provides best practice guidelines for applying multi-agent systems in domain-specific problem-solving and implementing effective agent tuning strategies. Our empirical studies, particularly in the financial question-answering domain, demonstrate that our approach achieves 95.0% of GPT-4's performance, while effectively managing costs and ensuring data privacy.

cs.AI↗

arXiv2024

PICO: Accelerating All k-Core Paradigms on GPU

Core decomposition is a well-established graph mining problem with various applications that involves partitioning the graph into hierarchical subgraphs. Solutions to this problem have been developed using both bottom-up and top-down approaches from the perspective of vertex convergence dependency. However, existing algorithms have not effectively harnessed GPU performance to expedite core decomposition, despite the growing need for enhanced performance. Moreover, approaching performance limitations of core decomposition from two different directions within a parallel synchronization structure has not been thoroughly explored. This paper introduces an efficient GPU acceleration framework, PICO, for the Peel and Index2core paradigms of k-core decomposition. We propose PeelOne, a Peel-based algorithm designed to simplify the parallel logic and minimize atomic operations by eliminating vertices that are 'under-core'. We also propose an Index2core-based algorithm, named HistoCore, which addresses the issue of extensive redundant computations across both vertices and edges. Extensive experiments on NVIDIA RTX 3090 GPU show that PeelOne outperforms all other Peel-based algorithms, and HistoCore outperforms all other Index2core-based algorithms. Furthermore, HistoCore even outperforms PeelOne by 1.1x - 3.2x speedup on six datasets, which breaks the stereotype that the Index2core paradigm performs much worse than the Peel in a shared memory parallel setting.

cs.DC↗

arXiv2023

Opto-twistronic Hall effect in a three-dimensional spiral lattice

Studies of moire systems have elucidated the exquisite effect of quantum geometry on the electronic bands and their properties, leading to the discovery of new correlated phases. However, most experimental studies have been confined to a few layers in the 2D limit. The extension of twistronics to its 3D limit, where the twist is extended into the third dimension between adjacent layers, remains underexplored due to the challenges in precisely stacking layers. Here, we focus on 3D twistronics on a platform of self-assembled spiral superlattice of multilayered WS2. Our findings reveal an opto-twistronic Hall effect in the spiral superlattice. This mesoscopic response is an experimental manifestation of the noncommutative geometry that arises when translational symmetry is replaced by a non-symmorphic screw operation. We also discover signatures of altered laws of optical excitation, manifested as an unconventional photon momentum-lattice interaction owing to moire of moire modulations in the 3D twistronic system. Crucially, our findings mark the initial identification of higher-order quantum geometrical tensors in light-matter interactions. This breakthrough opens new avenues for designing quantum materials-based optical lattices with large nonlinearities, paving the way for the development of advanced quantum nanophotonic devices.

cond-mat.mes-hall↗

arXiv2023

High Q and high gradient performance of the first medium-temperature baking 1.3 GHz cryomodule

World's first 1.3 GHz cryomodule containing eight 9-cell superconducting radio-frequency (RF) cavities treated by medium-temperature furnace baking (mid-T bake) was developed, assembled and tested at IHEP for the Dalian Advanced Light Source (DALS) and CEPC R&D. The 9-cell cavities in the cryomodule achieved an unprecedented highest average Q0 of 3.8E10 at 16 MV/m and 3.6E10 at 21 MV/m in the horizontal test. The cryomodule can operate stably up to a total CW RF voltage greater than 191 MV, with an average cavity CW accelerating gradient of more than 23 MV/m. The results significantly exceed the specifications of CEPC, DALS and the other high repetition rate free electron laser facilities (LCLS-II, LCLS-II-HE, SHINE, S3FEL). There is evidence that the mid-T bake cavity may not require fast cool-down or long processing time in the cryomodule. This paper reviews the cryomodule performance and discusses some important issues in cryomodule assembly and testing.

physics.acc-ph↗

arXiv2022

Atomically Sharp Internal Interface in a Chiral Weyl Semimetal Nanowire

Internal interfaces in Weyl semimetals (WSMs) are predicted to host distinct topological features that are different from the commonly studied external interfaces (crystal-to-vacuum boundaries). However, the lack of atomically sharp and crystallographically oriented internal interfaces in WSMs makes it difficult to experimentally investigate hidden topological states buried inside the material. Here, we study a unique internal interface known as merohedral twin boundary in chemically synthesized single-crystal nanowires (NWs) of CoSi, a chiral WSM of space group P213 (No. 198). High resolution scanning transmission electron microscopy reveals that this internal interface is (001) twin plane and connects two enantiomeric counterparts at an atomically sharp interface with inversion twinning. Ab-initio calculations show localized internal Fermi arcs at the (001) twin boundary that can be clearly distinguished from both external Fermi arcs and bulk states. These merohedrally twinned CoSi NWs provide an ideal material system to probe unexplored topological properties associated with internal interfaces in WSMs.

cond-mat.mtrl-sci↗

arXiv2022

Transport signatures of Fermi arcs at twin boundaries in Weyl materials

One of the most striking signatures of Weyl fermions is their surface Fermi arcs. Less known is that Fermi arcs can also be localized at internal twin boundaries where two Weyl materials of opposite chirality meet. In this work, we derive constraints on the topology and connectivity of these "internal Fermi arcs." We show that internal Fermi arcs can exhibit transport signatures and propose two probes: quantum oscillations and a quantized chiral magnetic current. We propose merohedrally twinned B20 materials as candidates to host internal Fermi arcs, verified through both model and ab initio calculations. Our theoretical investigation sheds lights on the topological features and motivates experimental studies into the intriguing physics of internal Fermi arcs.

cond-mat.mes-hall↗

arXiv2022

Direction and Trajectory Tracking Control for Nonholonomic Spherical Robot by Combining Sliding Mode Controller and Model Prediction Controller

Spherical robot is a nonlinear, nonholonomic and unstable system which increases the difficulty of the direction and trajectory tracking problem. In this study, we propose a new direction controller HTSMC, an instruction planning controller MPC, and a trajectory tracking framework MHH. The HTSMC is designed by integrating a fast terminal algorithm, a hierarchical method, the motion features of a spherical robot, and its dynamics. In addition, the new direction controller has an excellent control effect with a quick response speed and strong stability. MPC can obtain optimal commands that are then transmitted to the velocity and direction controller. Since the two torque controllers in MHH are all Lyapunov-based sliding mode controllers, the MHH framework may achieve optimal control performance while assuring stability. Finally, the two controllers eliminate the requirement for MPC's stability and dynamic constraints. Finally, hardware experiments demonstrate the efficacy of the HTSMC, MPC, and MHH.

cs.RO↗

arXiv2022

Disentangling magnetic and grain contrast in polycrystalline FeGe thin films using four-dimensional Lorentz scanning transmission electron microscopy

The study of nanoscale chiral magnetic order in polycrystalline materials with a strong Dzyaloshinkii-Moriya interaction (DMI) is interesting for the observation of magnetic phenomena at grain boundaries and interfaces. One such material is sputter-deposited B20 FeGe on Si, which has been actively investigated as the basis for low-power, high-density magnetic memory technology in a scalable material platform. Although conventional Lorentz electron microscopy provides the requisite spatial resolution to probe chiral magnetic textures in single-crystal FeGe, probing the magnetism of sputtered B20 FeGe is more challenging because the sub-micron crystal grains add confounding contrast. We address the challenge of disentangling magnetic and grain contrast by applying 4-dimensional Lorentz scanning transmission electron microscopy using an electron microscope pixel array detector. Supported by analytical and numerical models, we find that the most important parameter for imaging magnetic materials with polycrystalline grains is the ability for the detector to sustain large electron doses, where having a high-dynamic range detector becomes extremely important. Despite the small grain size in sputtered B20 FeGe on Si, using this approach we are still able to observe helicity switching of skyrmions and magnetic helices across two adjacent grains as they thread through neighboring grains. We reproduce this effect using micromagnetic simulations by assuming that the grains have distinct orientation and magnetic chirality and find that magnetic helicity couples to crystal chirality. Our methodology for imaging magnetic textures is applicable to other thin-film magnets used for spintronics and memory applications, where an understanding of how magnetic order is accommodated in polycrystalline materials is important.

cond-mat.mtrl-sci↗

arXiv2020

Medium-temperature furnace bake of Superconducting Radio-Frequency cavities at IHEP

Recently, heat treatment between 250 C and 500 C has been attempted to improve quality factor of superconducting radio-frequency cavities at FNAL and KEK. Experiments of such medium temperature (mid-T) bake with furnaces have also been carried out at IHEP. Firstly, eleven 1.3 GHz 1-cell cavities were treated with different temperatures at a small furnace. The average quality factor has reached 3.6E10 when the gradient is 16 MV/m. Then, the recipe of mid-T furnace bake at 300 C for 3 hours has been applied to six 1.3 GHz 9-cell cavities at a new big furnace. The average quality factor has reached 3.8E10 when the gradient is 16 MV/m.

physics.acc-ph↗

arXiv2020

Automatic acute ischemic stroke lesion segmentation using semi-supervised learning

Ischemic stroke is a common disease in the elderly population, which can cause long-term disability and even death. However, the time window for treatment of ischemic stroke in its acute stage is very short. To fast localize and quantitively evaluate the acute ischemic stroke (AIS) lesions, many deep-learning-based lesion segmentation methods have been proposed in the literature, where a deep convolutional neural network (CNN) was trained on hundreds of fully labeled subjects with accurate annotations of AIS lesions. Despite that high segmentation accuracy can be achieved, the accurate labels should be annotated by experienced clinicians, and it is therefore very time-consuming to obtain a large number of fully labeled subjects. In this paper, we propose a semi-supervised method to automatically segment AIS lesions in diffusion weighted images and apparent diffusion coefficient maps. By using a large number of weakly labeled subjects and a small number of fully labeled subjects, our proposed method is able to accurately detect and segment the AIS lesions. In particular, our proposed method consists of three parts: 1) a double-path classification net (DPC-Net) trained in a weakly-supervised way is used to detect the suspicious regions of AIS lesions; 2) a pixel-level K-Means clustering algorithm is used to identify the hyperintensive regions on the DWIs; and 3) a region-growing algorithm combines the outputs of the DPC-Net and the K-Means to obtain the final precise lesion segmentation. In our experiment, we use 460 weakly labeled subjects and 15 fully labeled subjects to train and fine-tune the proposed method. By evaluating on a clinical dataset with 150 fully labeled subjects, our proposed method achieves a mean dice coefficient of 0.642, and a lesion-wise F1 score of 0.822.

eess.IV↗

arXiv2020

Quantum interference between the optical Stark effect and resonant harmonic generation in WS2

An applied field can modulate optical signals by resonance shifting via the Stark effect. The optical Stark effect (OSE) uses ultrafast light in the transparency region of a material to shift resonances with speeds limited by the pulse duration or system coherence. In this Letter we investigate the OSE in resonant optical harmonic generation (OHG) using the ground state exciton transition of WS2 with a variety of morphologies. Multidimensional pump-harmonic-probe measurements, in which the probe is second- or third-harmonic emission, reveal not only large Stark shifts that are commensurate with the large optical susceptibilities common to WS2 excitons, but also behaviors more complex than simple OSE treatments predict. We show how a new manifestation of the Stark Effect, brought forth by coherent photon exchange between the pump and OHG fundamental fields, can strongly enhance or suppress OHG.

physics.optics↗