arXiv subjects

Kang He

Publications and source records attributed to Kang He.

31 records · Page 2Linked to original sources

arXiv2026

LongCLI-Bench: A Preliminary Benchmark and Study for Long-horizon Agentic Programming in Command-Line Interfaces

Recent advances in AI-assisted programming have empowered agents to execute complex workflows via command-line interfaces, however, existing benchmarks are limited by short task horizons, data contamination from GitHub scraping, and a lack of fine-grained evaluation metrics, fail to rigorously evaluate the long-horizon planning and execution capabilities essential for realistic software engineering. To address these gaps, we introduce LongCLI-Bench, a comprehensive benchmark designed to evaluate agentic capabilities across long-horizon, realistic tasks. We curated 20 high-quality, long-horizon tasks from over 1,000 computer science assignments and real-world workflows, covering four engineering categories: from scratch, feature addition, bug fixing, and refactoring. We propose a dual-set testing protocol for LongCLI-Bench, which measures requirement fulfillment (fail-to-pass) and regression avoidance (pass-to-pass), and incorporates step-level scoring to pinpoint execution failures. Extensive experiments reveal that even state-of-the-art agents achieve pass rates below 20% in LongCLI-Bench. Step-level analysis further indicates that the majority of tasks stall at less than 30% completion, highlighting that critical failures often occur in the early stages. Although self-correction offers marginal gains, human-agent collaboration through plan injection and interactive guidance yields significantly higher improvements. These results highlight that future research must emphasize the development of synergistic human-agent workflows alongside advances in agents' planning and execution capabilities to overcome key challenges in long-horizon task performance.

cs.SE↗

arXiv2025

Kling-Omni Technical Report

We present Kling-Omni, a generalist generative framework designed to synthesize high-fidelity videos directly from multimodal visual language inputs. Adopting an end-to-end perspective, Kling-Omni bridges the functional separation among diverse video generation, editing, and intelligent reasoning tasks, integrating them into a holistic system. Unlike disjointed pipeline approaches, Kling-Omni supports a diverse range of user inputs, including text instructions, reference images, and video contexts, processing them into a unified multimodal representation to deliver cinematic-quality and highly-intelligent video content creation. To support these capabilities, we constructed a comprehensive data system that serves as the foundation for multimodal video creation. The framework is further empowered by efficient large-scale pre-training strategies and infrastructure optimizations for inference. Comprehensive evaluations reveal that Kling-Omni demonstrates exceptional capabilities in in-context generation, reasoning-based editing, and multimodal instruction following. Moving beyond a content creation tool, we believe Kling-Omni is a pivotal advancement toward multimodal world simulators capable of perceiving, reasoning, generating and interacting with the dynamic and complex worlds.

cs.CV↗

arXiv2025

KlingAvatar 2.0 Technical Report

Avatar video generation models have achieved remarkable progress in recent years. However, prior work exhibits limited efficiency in generating long-duration high-resolution videos, suffering from temporal drifting, quality degradation, and weak prompt following as video length increases. To address these challenges, we propose KlingAvatar 2.0, a spatio-temporal cascade framework that performs upscaling in both spatial resolution and temporal dimension. The framework first generates low-resolution blueprint video keyframes that capture global semantics and motion, and then refines them into high-resolution, temporally coherent sub-clips using a first-last frame strategy, while retaining smooth temporal transitions in long-form videos. To enhance cross-modal instruction fusion and alignment in extended videos, we introduce a Co-Reasoning Director composed of three modality-specific large language model (LLM) experts. These experts reason about modality priorities and infer underlying user intent, converting inputs into detailed storylines through multi-turn dialogue. A Negative Director further refines negative prompts to improve instruction alignment. Building on these components, we extend the framework to support ID-specific multi-character control. Extensive experiments demonstrate that our model effectively addresses the challenges of efficient, multimodally aligned long-form high-resolution video generation, delivering enhanced visual clarity, realistic lip-teeth rendering with accurate lip synchronization, strong identity preservation, and coherent multimodal instruction following.

cs.CV↗

arXiv2025

PaSE: Prototype-aligned Calibration and Shapley-based Equilibrium for Multimodal Sentiment Analysis

Multimodal Sentiment Analysis (MSA) seeks to understand human emotions by integrating textual, acoustic, and visual signals. Although multimodal fusion is designed to leverage cross-modal complementarity, real-world scenarios often exhibit modality competition: dominant modalities tend to overshadow weaker ones, leading to suboptimal performance. In this paper, we propose PaSE, a novel Prototype-aligned Calibration and Shapley-optimized Equilibrium framework, which enhances collaboration while explicitly mitigating modality competition. PaSE first applies Prototype-guided Calibration Learning (PCL) to refine unimodal representations and align them through an Entropic Optimal Transport mechanism that ensures semantic consistency. To further stabilize optimization, we introduce a Dual-Phase Optimization strategy. A prototype-gated fusion module is first used to extract shared representations, followed by Shapley-based Gradient Modulation (SGM), which adaptively adjusts gradients according to the contribution of each modality. Extensive experiments on IEMOCAP, MOSI, and MOSEI confirm that PaSE achieves the superior performance and effectively alleviates modality competition.

cs.LG↗

arXiv2025

P-orbital spin generator with large spin Hall angle and long spin diffusion length

High density data storage and spin-logic devices require highly efficient all-electric control of spin moments. So far, charge-to-spin conversion through the spin Hall effect (SHE) highly limits to d-orbital materials associated with strong spin-orbit coupling (SOC), especially heavy metals. However, d-orbital heavy metals with strong SOC results in a short spin diffusion length, which restricts the spin transport and accumulation in spintronic devices. Therefore, it is urgent to discovery new SHE materials with both large spin Hall conductivity and high spin transport ability beyond d-orbital materials. Here, we experimentally report a large charge to spin conversion in a p-orbital In2Bi alloy, exhibiting the coexistence of a large spin Hall angle and a long spin diffusion length (4 times that of Pt). Our first-principles calculations reveal that small gap openings near the Fermi level lead to large Berry curvature-related spin Hall conductivity. Due to the delocalized nature of p-orbitals of In2Bi, its spin current can overcome the physical barriers between spin Hall angle and spin diffusion length in d-orbital metals, thereby advancing the development of high performance spintronic devices.

cond-mat.mtrl-sci↗

arXiv2025

LogicTree: Structured Proof Exploration for Coherent and Rigorous Logical Reasoning with Large Language Models

Large language models (LLMs) have achieved remarkable multi-step reasoning capabilities across various domains. However, LLMs still face distinct challenges in complex logical reasoning, as (1) proof-finding requires systematic exploration and the maintenance of logical coherence and (2) searching the right combination of premises at each reasoning step is inherently challenging in tasks with large premise space. To address this, we propose LogicTree, an inference-time modular framework employing algorithm-guided search to automate structured proof exploration and ensure logical coherence. Advancing beyond tree-of-thought (ToT), we incorporate caching mechanism into LogicTree to enable effective utilization of historical knowledge, preventing reasoning stagnation and minimizing redundancy. Furthermore, we address the combinatorial complexity of premise search by decomposing it into a linear process. The refined premise selection restricts subsequent inference to at most one derivation per step, enhancing reasoning granularity and enforcing strict step-by-step reasoning. Additionally, we introduce two LLM-free heuristics for premise prioritization, enabling strategic proof search. Experimental results on five datasets demonstrate that LogicTree optimally scales inference-time computation to achieve higher proof accuracy, surpassing chain-of-thought (CoT) and ToT with average gains of 23.6% and 12.5%, respectively, on GPT-4o. Moreover, within LogicTree, GPT-4o outperforms o3-mini by 7.6% on average.

cs.CL↗

arXiv2025

DALR: Dual-level Alignment Learning for Multimodal Sentence Representation Learning

Previous multimodal sentence representation learning methods have achieved impressive performance. However, most approaches focus on aligning images and text at a coarse level, facing two critical challenges:cross-modal misalignment bias and intra-modal semantic divergence, which significantly degrade sentence representation quality. To address these challenges, we propose DALR (Dual-level Alignment Learning for Multimodal Sentence Representation). For cross-modal alignment, we propose a consistency learning module that softens negative samples and utilizes semantic similarity from an auxiliary task to achieve fine-grained cross-modal alignment. Additionally, we contend that sentence relationships go beyond binary positive-negative labels, exhibiting a more intricate ranking structure. To better capture these relationships and enhance representation quality, we integrate ranking distillation with global intra-modal alignment learning. Comprehensive experiments on semantic textual similarity (STS) and transfer (TR) tasks validate the effectiveness of our approach, consistently demonstrating its superiority over state-of-the-art baselines.

cs.CL↗

arXiv2025

Zero-Shot Conversational Stance Detection: Dataset and Approaches

Stance detection, which aims to identify public opinion towards specific targets using social media data, is an important yet challenging task. With the increasing number of online debates among social media users, conversational stance detection has become a crucial research area. However, existing conversational stance detection datasets are restricted to a limited set of specific targets, which constrains the effectiveness of stance detection models when encountering a large number of unseen targets in real-world applications. To bridge this gap, we manually curate a large-scale, high-quality zero-shot conversational stance detection dataset, named ZS-CSD, comprising 280 targets across two distinct target types. Leveraging the ZS-CSD dataset, we propose SITPCL, a speaker interaction and target-aware prototypical contrastive learning model, and establish the benchmark performance in the zero-shot setting. Experimental results demonstrate that our proposed SITPCL model achieves state-of-the-art performance in zero-shot conversational stance detection. Notably, the SITPCL model attains only an F1-macro score of 43.81%, highlighting the persistent challenges in zero-shot conversational stance detection.

cs.CL↗

arXiv2024

Prompt-Based Bias Calibration for Better Zero/Few-Shot Learning of Language Models

Prompt-based learning is susceptible to intrinsic bias present in pre-trained language models (LMs), leading to sub-optimal performance in prompt-based zero/few-shot settings. In this work, we propose a null-input prompting method to calibrate intrinsic bias encoded in pre-trained LMs. Different from prior efforts that address intrinsic bias primarily for social fairness and often involve excessive computational cost, our objective is to explore enhancing LMs' performance in downstream zero/few-shot learning while emphasizing the efficiency of intrinsic bias calibration. Specifically, we leverage a diverse set of auto-selected null-meaning inputs generated from GPT-4 to probe intrinsic bias of pre-trained LMs. Utilizing the bias-reflected probability distribution, we formulate a distribution disparity loss for bias calibration, where we exclusively update bias parameters ($0.1\%$ of total parameters) of LMs towards equal probability distribution. Experimental results show that the calibration promotes an equitable starting point for LMs while preserving language modeling abilities. Across a wide range of datasets, including sentiment analysis and topic classification, our method significantly improves zero/few-shot learning performance of LMs for both in-context learning and prompt-based fine-tuning (on average $9\%$ and $2\%$, respectively).

cs.CL↗

arXiv2024

Revisit to the WGVC schemes: a nonlinear order-preserving and spectral-property-optimized methodology and its enhancement

The numerical simulation of supersonic complex flow problems demands capabilities in identifying multiscale structures and capturing shocks, imposing stringent requirements on the numerical scheme. The capability to identify multiscale structures is closely related to the spectral properties of the numerical scheme. Currently, existing methods to improve the spectral properties of finite difference schemes face shortcomings such as parallel difficulties (compact methods) or introducing unnecessary dispersion errors at low wavenumbers due to accuracy loss (spectral-like optimization methods). In this paper, we proposed an order-preserving spectral properties optimization method based on the group velocity control theory: the weighted group velocity control (WGVC) scheme. This method, centered around the concept of group velocity, achieves low-wavenumber accuracy control and mid-wavenumber group velocity control by designing smoothness indicators and nonlinear weighting approach for wave packets. Furthermore, by embedding the WGVC scheme into shock-capturing schemes such as WENO/TENO scheme, we not only preserve the spectral properties of the WGVC scheme at medium to low wavenumbers but also enhance the shock-capturing capability of the scheme. Theoretical and numerical experiments verify that the new method has advantages such as order-preserving, small dispersion and dissipation errors, and is very suitable for numerical simulation of complex flow problems such as turbulence-shock boundary layer interactions.

physics.flu-dyn↗

arXiv2022

Planar Hall effect induced spin rectification effect and its strong impact on spin pumping measurements

Spin pumping is a technique widely used to generate the pure spin current and characterize the spin-charge conversion in various systems. The reversing sign of the symmetric Lorentzian charge current with respect to opposite magnetic field is generally accepted as the key criterion to identify its pure spin current origin. However, we herein find that the rectified voltage due to the planar Hall effect can exhibit similar spurious signal, complicating and even misleading the analysis. The distribution of microwave magnetic field and induction current has strong influence on the magnetic field symmetry and lineshape of the obtained signal. We further demonstrate a geometry where the spin-charge conversion and the rectified voltage can be readily distinguished with a straightforward symmetry analysis.

cond-mat.mes-hall↗

arXiv2020

Gate-Tunable Semiconductor Heterojunctions from 2D/3D van der Waals Interfaces

Van der Waals (vdW) semiconductors are attractive for highly scaled devices and heterogeneous integration since they can be isolated into self-passivated, two-dimensional (2D) layers that enable superior electrostatic control. These attributes have led to numerous demonstrations of field-effect devices ranging from transistors to triodes. By exploiting the controlled, substitutional doping schemes in covalently-bonded, three-dimensional (3D) semiconductors and the passivated surfaces of 2D semiconductors, one can construct devices that can exceed performance metrics of 'all-2D' vdW heterojunctions. Here, we demonstrate, 2D/3D semiconductor heterojunctions using MoS2 as the prototypical 2D semiconductor laid upon Si and GaN as the 3D semiconductor layers. By tuning the Fermi levels in MoS2, we demonstrate devices that concurrently exhibit over seven orders of magnitude modulation in rectification ratios and conductance. Our results further suggest that the interface quality does not necessarily affect Fermi-level tuning at the junction opening up possibilities for novel 2D/3D heterojunction device architectures.

cond-mat.mes-hall↗

arXiv2020

Ghost spintronic THz-emitter-array microscope

Terahertz (THz) wave shows great potential in non-destructive testing, bio detection and cancer imaging. Recent progresses on THz wave near-field probes/apertures enable mechanically raster scanning of an object's surface in the near-field region, while an efficient, non-scanning, non-invasive, deeply sub-diffraction-limited imaging still remains challenging. Here, we demonstrate a THz near-field microscopy using a reconfigurable spintronic THz emitter array (STEA) with computational ghost imaging. By illuminating an object with the reconfigurable STEA in near field and computing the correlation measurements, we reconstruct its image with deeply sub-diffraction resolution. By circulating an external magnetic field, the in-line polarization rotation of THz waves is realized, making the fused image contrast polarization-free. The time-of-flight (TOF) measurements of coherent THz pulses further enables to resolve objects at different distances or depths. The demonstrated ghost spintronic THz-emitter-array microscope (GHOSTEAM) is a new imaging tool for THz near-field real-time imaging (with potential of video framerate), especially opening up paradigm-shift opportunities in non-intrusive label-free bioimaging in a broadband frequency range from 0.1 THz to 30 THz (namely 3.3-1000 cm-1).

physics.optics↗