arXiv subjects

Haoyan Yang

Publications and source records attributed to Haoyan Yang.

17 recordsLinked to original sources

arXiv2026

The Research and Development of New Electronics System and its Testing on the JNE-1ton Prototype Detector

The Jinping Neutrino Experiment (JNE), a next-generation neutrino observatory under construction at the China Jinping Underground Laboratory II (CJPL-II), requires high-precision waveform-based event reconstruction, imposing stringent demands on its readout electronics. To meet these requirements, we have developed a high-performance readout system featuring 1 GSa/s real-time sampling, 14-bit physical resolution with an effective number of bits (ENOB) of 10.6, a total data throughput of 64 Gbps, and a deterministic zero-delay clock distribution architecture. The new single-crate 64-channel system (PDS1500) was validated through bench tests and deployment on the upgraded JNE-1ton prototype detector. Its performance was further evaluated against a commercial reference system. The results demonstrate that all key metrics meet the JNE experimental requirements: zero data loss within a 1000 ns acquisition window, baseline noise reduced to one-third of the reference level, timing drift limited to 0.3 ns across power cycles, and an energy threshold as low as 0.1 MeV, enabling the detection of low-energy solar neutrinos. While the 14-bit physical resolution provides significantly higher waveform fidelity, the overall energy resolution in this test remains dominated by the intrinsic limitations of the JNE-1ton detector, as expected. Furthermore, the modular architecture provides the throughput and scalability required to support the full-scale 3000-channel JNE detector. These results collectively demonstrate that the newly developed electronics system fully satisfies the technical requirements of the future JNE experiment.

physics.ins-det↗

arXiv2026

Mountain Muography for China Jinping Underground Laboratory

The China Jinping Underground Laboratory (CJPL), located $\sim 2,400$~m beneath Jinping Mountain, is one of the world's deepest and largest ($\sim 300{,}000~\mathrm{m}^3$) underground facilities, hosting dark matter, nuclear astrophysics, and neutrino experiments. We report the first muon radiography (muography) conducted at this extraordinary depth. Cosmic muons detected by a one-ton prototype developed for the Jinping Neutrino Experiment were used to perform non-invasive subsurface density mapping over a 3~km lateral range. The 1.3~m diameter detector provides nearly isotropic acceptance and an angular resolution of $\sim 4.5^\circ$. By correlating the predicted surface muon flux distributions with the underground measurements, we reconstruct a directional opacity map that constrains the density structure of the overburden and shows excellent agreement with satellite-derived terrain models. This work demonstrates the feasibility of muography at extreme depths with kilometer-scale overburden and establishes a robust methodology for future geophysical applications and large-scale facilities, such as the full Jinping Neutrino Experiment. Based on this validated overburden model, we further predict the total muon fluxes for the eight experimental halls in CJPL-II, providing essential input for their physics programs.

hep-ex↗

arXiv2026

Capability Self-Assessment: Teaching LLMs to Know Their Limits

The ability to recognize one's own limitations and decide whether to solve a problem or delegate is fundamental for reliable intelligent systems. Yet we show that modern large language models systematically lack this ability: across diverse model families and scales, they overestimate their competence and attempt queries they cannot solve. We refer to this ability as Capability Self-Assessment (CSA) and formulate it as a policy-learning problem, aiming to improve self-assessment while preserving the model's original capabilities. Our results show that reinforcement learning teaches CSA effectively, significantly outperforming supervised fine-tuning while preserving original capabilities. In contrast, supervised fine-tuning severely degrades the capabilities the model is meant to assess. Moreover, learned self-assessment behavior generalizes well out of distribution, suggesting that CSA is a transferable model trait. Finally, CSA is practically useful: it improves local-cloud decision making at inference time and provides a signal for targeted data selection during training.

cs.AI↗

arXiv2026

SkillsVote: Lifecycle Governance of Agent Skills from Collection, Recommendation to Evolution

Long-horizon LLM agents generate traces that could become reusable experience, but raw trajectories are noisy, local, and hard to govern. Agent Skills offer a structured artifact for combining procedural guidance, executable resources, and applicability boundaries. Yet open skill ecosystems contain redundant, uneven, environment-sensitive artifacts, and indiscriminate updates can pollute future context. We present SkillsVote, a lifecycle-governance framework for Agent Skills across collection, recommendation, attribution, and evolution. SkillsVote profiles a million-scale open source corpus for environment requirements, quality, and verifiability, and synthesizes tasks for verifiable skills. Before execution, it performs agentic library search over structured skill folders to expose instructional context. After execution, it decomposes trajectories into skill-linked subtasks, attributes outcomes to skill-guided execution, agent exploration, environment, and result signals, and admits only successful reusable discoveries to evidence-gated updates. Experiments on Terminal-Bench 2.0 and SWE-Bench Pro show that SkillsVote improves agent performance on challenging agentic coding benchmarks. The gains arise from two complementary pathways: online evolution over task streams at test time and offline transfer via frozen libraries built from either historical trajectories or curated open source skills.

cs.CL↗

arXiv2026

External Validation of Deep Learning Models for BI-RADS Breast Density Prediction from Ultrasound Images

We externally validated three deep learning models (DenseNet121, ViT-B/32, and ResNet50) for predicting mammographic breast density from breast ultrasound exams on an independent cohort. The external validation set comprised 2,000 ultrasound exams, including 500 cancer cases defined by an initial negative exam (BI-RADS 1 or 2) followed by a cancer diagnosis within 6 months to 10 years, and 1,500 negative controls matched by manufacturer and study year. Performance was measured using patient-level AUROC across four density categories: A (fatty), B (scattered), C (heterogeneous), and D (extremely dense). As a downstream assessment, we also evaluated 10-year risk prediction by incorporating age and AI-derived density into the Tyrer-Cuzick model and comparing performance against a reference model using age and mammography-reported density. All three models performed best in extremely dense breasts (AUROC 0.868-0.899), with strong performance in fatty (0.814-0.838) and scattered density (0.764-0.799), and lower performance in heterogeneously dense breasts (0.699-0.729). DenseNet121 achieved the highest overall performance (micro-averaged AUROC 0.885), and performance across categories was comparable between internal and external testing. For risk modeling, age combined with AI-derived density yielded a lower AUROC than age combined with mammography-reported density (0.541 vs. 0.570; p = 0.23), with no statistically significant difference. These findings indicate that deep learning models generalize well to external data with different racial composition for breast density assessment. While performance is strongest in extremely dense breasts, heterogeneously dense remains more challenging, highlighting the need for targeted optimization.

eess.IV↗

arXiv2026

Self-Improvement of Large Language Models: A Technical Overview and Future Outlook

As large language models (LLMs) continue to advance, improving them solely through human supervision is becoming increasingly costly and limited in scalability. As models approach human-level capabilities in certain domains, human feedback may no longer provide sufficiently informative signals for further improvement. At the same time, the growing ability of models to make autonomous decisions and execute complex actions naturally enables abstractions in which components of the model development process can be progressively automated. Together, these challenges and opportunities have driven increasing interest in self-improvement, where models autonomously generate data, evaluate outputs, and iteratively refine their own capabilities. In this paper, we present a system-level perspective on self-improving language models and introduce a unified framework that organizes existing techniques. We conceptualize the self-improvement system as a closed-loop lifecycle, consisting of four tightly coupled processes: data acquisition, data selection, model optimization, and inference refinement, along with an autonomous evaluation layer throughout the process. Within this framework, the model itself plays a central role in driving each stage: collecting or generating data, selecting informative signals, updating its parameters, and refining outputs, while the autonomous evaluation layer continuously monitors progress and guides the improvement cycle across stages. Following this lifecycle perspective, we systematically review and analyze representative methods for each component from a technical standpoint. We further examine current limitations, potential risks, and prominent applications, and outline our vision for future research toward fully self-improving LLMs.

cs.CL↗

arXiv2025

Investigating Production of TeV-scale Muons in Extensive Air Shower at 2400 Meters Underground

Deep underground experiments present a new avenue to probe the first interactions in extensive air showers or hadronic interactions in the extreme forward phase space. The China Jinping Underground Laboratory, characterized by a vertical rock overburden of 2,400~m, provides an exceptionally effective shield against cosmic muons with energies below 3~TeV. The surviving high-energy muons, produced in the first interactions of extensive air showers, open a unique observational window into primary cosmic rays from tens of TeV up to the PeV scale and beyond. This distinctive feature also enables detailed studies of charged hadron production in the earliest stages of shower development. Using 1,338.6 live days of data collected with a one-ton prototype detector for the Jinping Neutrino Experiment, we measured the underground muon flux originating from air showers. The results show discrepancies of about 40\% corresponding to significances of more than 2$\sigma$, relative to predictions from several leading hadronic interaction models. We interpret these findings from two complementary perspectives: (i) by adopting the expected cosmic-ray spectra, we constrain the modeling of the first hadronic interactions in air showers and provide novel insights into resolving the long-standing \textit{muon puzzle}; and (ii) by assuming specific hadronic interaction models, we infer the mass composition of cosmic rays, and our data favor a lighter component in the corresponding energy range. Our study demonstrates the potential of deep underground laboratories to provide new experimental insights into air shower physics and cosmic rays.

hep-ex↗

arXiv2025

Any Large Language Model Can Be a Reliable Judge: Debiasing with a Reasoning-based Bias Detector

LLM-as-a-Judge has emerged as a promising tool for automatically evaluating generated outputs, but its reliability is often undermined by potential biases in judgment. Existing efforts to mitigate these biases face key limitations: in-context learning-based methods fail to address rooted biases due to the evaluator's limited capacity for self-reflection, whereas fine-tuning is not applicable to all evaluator types, especially closed-source models. To address this challenge, we introduce the Reasoning-based Bias Detector (RBD), which is a plug-in module that identifies biased evaluations and generates structured reasoning to guide evaluator self-correction. Rather than modifying the evaluator itself, RBD operates externally and engages in an iterative process of bias detection and feedback-driven revision. To support its development, we design a complete pipeline consisting of biased dataset construction, supervision collection, distilled reasoning-based fine-tuning of RBD, and integration with LLM evaluators. We fine-tune four sizes of RBD models, ranging from 1.5B to 14B, and observe consistent performance improvements across all scales. Experimental results on 4 bias types--verbosity, position, bandwagon, and sentiment--evaluated using 8 LLM evaluators demonstrate RBD's strong effectiveness. For example, the RBD-8B model improves evaluation accuracy by an average of 18.5% and consistency by 10.9%, and surpasses prompting-based baselines and fine-tuned judges by 12.8% and 17.2%, respectively. These results highlight RBD's effectiveness and scalability. Additional experiments further demonstrate its strong generalization across biases and domains, as well as its efficiency.

cs.CL↗

arXiv2025

Development and characterization of the JNE concentrator

The Jinping Neutrino Experiment (JNE) will utilize approximately 3000 8-inch MCP-PMTs identified as GDB-6082 from North Night Vision Technology to detect neutrinos. To enhance the effective coverage of the JNE detector, mounting a custom-designed light concentrator on each PMT is a practical and economical approach. We measured angular responses of the concentration factor at four wavelengths in air medium for the concentrator with the selected cutoff angle of 70{\deg}. The measurements align with the Monte Carlo simulations. Furthermore, our results indicate that these concentrators can improve the efficiency of light collection by 40 % under parallel illumination conditions. This enhancement results in a slight increase in transit-time spread, with the full width at half maximum (FWHM) increasing by less than 0.3 ns. We conclude that the developed light concentrators are highly suitable for the JNE.

hep-ex↗

arXiv2025

Dynamic Noise Preference Optimization: Self-Improvement of Large Language Models with Self-Synthetic Data

Although LLMs have achieved significant success, their reliance on large volumes of human-annotated data has limited their potential for further scaling. In this situation, utilizing self-generated synthetic data has become crucial for fine-tuning LLMs without extensive human annotation. However, current methods often fail to ensure consistent improvements across iterations, with performance stagnating after only minimal updates. To overcome these challenges, we introduce Dynamic Noise Preference Optimization (DNPO), which combines dynamic sample labeling for constructing preference pairs with controlled, trainable noise injection during preference optimization. Our approach effectively prevents stagnation and enables continuous improvement. In experiments with Llama-3.2-3B and Zephyr-7B, DNPO consistently outperforms existing methods across multiple benchmarks. Additionally, with Zephyr-7B, DNPO shows a significant improvement in model-generated data quality, with a 29.4% win-loss rate gap compared to the baseline in GPT-4 evaluations.

cs.CL↗

arXiv2024

Exploring Performance Contrasts in TableQA: Step-by-Step Reasoning Boosts Bigger Language Models, Limits Smaller Language Models

This paper proposes a detailed prompting flow, termed Table-Logic, to investigate the performance contrasts between bigger and smaller language models (LMs) utilizing step-by-step reasoning methods in the TableQA task. The method processes tasks by sequentially identifying critical columns and rows given question and table with its structure, determining necessary aggregations, calculations, or comparisons, and finally inferring the results to generate a precise prediction. By deploying this method, we observe a 7.8% accuracy improvement in bigger LMs like Llama-3-70B compared to the vanilla on HybridQA, while smaller LMs like Llama-2-7B shows an 11% performance decline. We empirically investigate the potential causes of performance contrasts by exploring the capabilities of bigger and smaller LMs from various dimensions in TableQA task. Our findings highlight the limitations of the step-by-step reasoning method in small models and provide potential insights for making improvements.

cs.CL↗

arXiv2024

BURExtract-Llama: An LLM for Clinical Concept Extraction in Breast Ultrasound Reports

Breast ultrasound is essential for detecting and diagnosing abnormalities, with radiology reports summarizing key findings like lesion characteristics and malignancy assessments. Extracting this critical information is challenging due to the unstructured nature of these reports, with varied linguistic styles and inconsistent formatting. While proprietary LLMs like GPT-4 are effective, they are costly and raise privacy concerns when handling protected health information. This study presents a pipeline for developing an in-house LLM to extract clinical information from radiology reports. We first use GPT-4 to create a small labeled dataset, then fine-tune a Llama3-8B model on it. Evaluated on clinician-annotated reports, our model achieves an average F1 score of 84.6%, which is on par with GPT-4. Our findings demonstrate the feasibility of developing an in-house LLM that not only matches GPT-4's performance but also offers cost reductions and enhanced data privacy.

cs.CL↗

arXiv2024

PFID: Privacy First Inference Delegation Framework for LLMs

This paper introduces a novel privacy-preservation framework named PFID for LLMs that addresses critical privacy concerns by localizing user data through model sharding and singular value decomposition. When users are interacting with LLM systems, their prompts could be subject to being exposed to eavesdroppers within or outside LLM system providers who are interested in collecting users' input. In this work, we proposed a framework to camouflage user input, so as to alleviate privacy issues. Our framework proposes to place model shards on the client and the public server, we sent compressed hidden states instead of prompts to and from servers. Clients have held back information that can re-privatized the hidden states so that overall system performance is comparable to traditional LLMs services. Our framework was designed to be communication efficient, computation can be delegated to the local client so that the server's computation burden can be lightened. We conduct extensive experiments on machine translation tasks to verify our framework's performance.

cs.CL↗

arXiv2024

Can We Trust LLMs? Mitigate Overconfidence Bias in LLMs through Knowledge Transfer

The study explores mitigating overconfidence bias in LLMs to improve their reliability. We introduce a knowledge transfer (KT) method utilizing chain of thoughts, where "big" LLMs impart knowledge to "small" LLMs via detailed, sequential reasoning paths. This method uses advanced reasoning of larger models to fine-tune smaller models, enabling them to produce more accurate predictions with calibrated confidence. Experimental evaluation using multiple-choice questions and sentiment analysis across diverse datasets demonstrated the KT method's superiority over the vanilla and question-answer pair (QA) fine-tuning methods. The most significant improvement in three key metrics, where the KT method outperformed the vanilla and QA methods by an average of 55.3% and 43.1%, respectively. These findings underscore the KT method's potential in enhancing model trustworthiness and accuracy, offering precise outputs with well-matched confidence levels across various contexts.

cs.CL↗

arXiv2023

Preliminary Design of Scalable Hardware Integrated Platform for LLRF Application

In this paper, the SHIP4LLRF (Scalable Hardware Integrated Platform for LLRF) based on 6U VPX-standard was designed preliminarily, which includes 6U mother board and two HPC FPGA mezzanine cards (FMCs). The ADC and DAC FMC is based on ADS54J60 from TI and LTC2000Y-16 form ADI, respectively. The system mother board is based on Xilinx Kintex UltraScale KU060, which also features 64-bit DDR4 SDRAM, QSFP and USB3.0 interfaces. Each FMC connector is assigned 58 pairs of LVDS standard IOs and 8 pairs of GTH high-speed serial lanes. Besides, the mother board is equipped with the self-developed ZYNQBee2 module based on ZYNQ7010 for slow control such as EPICS. All ADC or DAC raw data in each SHIP4LLEF is compressed loss-less without triggering and transmitted to the process board. A scalar quantization method which is in development is used for lossless compression of ADC raw data, the process board will decompress the ADC data and perform a digital algorithm to measure the amplitude and phase of the high frequency signal. This de-sign is scalable for testing and upgradability, mean-while, the trigger-less data transmission enable this system participate in both local (rack-scale) and accelerator-wide communication networks.

cs.AR↗

arXiv2023

PRCA: Fitting Black-Box Large Language Models for Retrieval Question Answering via Pluggable Reward-Driven Contextual Adapter

The Retrieval Question Answering (ReQA) task employs the retrieval-augmented framework, composed of a retriever and generator. The generator formulates the answer based on the documents retrieved by the retriever. Incorporating Large Language Models (LLMs) as generators is beneficial due to their advanced QA capabilities, but they are typically too large to be fine-tuned with budget constraints while some of them are only accessible via APIs. To tackle this issue and further improve ReQA performance, we propose a trainable Pluggable Reward-Driven Contextual Adapter (PRCA), keeping the generator as a black box. Positioned between the retriever and generator in a Pluggable manner, PRCA refines the retrieved information by operating in a token-autoregressive strategy via maximizing rewards of the reinforcement learning phase. Our experiments validate PRCA's effectiveness in enhancing ReQA performance on three datasets by up to 20% improvement to fit black-box LLMs into existing frameworks, demonstrating its considerable potential in the LLMs era.

cs.CL↗

arXiv2023

Recent Advances in Rate Control: From Optimisation to Implementation and Beyond

Video coding is a video compression technique that compresses the original video sequence to produce a smaller archive file or reduce the transmission bandwidth under constraints on the visual quality loss. Rate control (RC) plays a critical role in video coding. It can achieve stable stream output in practical applications, especially real-time video applications such as video conferencing or game live streaming. Most RC algorithms either directly or indirectly characterise the relationship between the bit rate (R) and quantisation (Q) and then allocate bits to every coding unit so as to guarantee the global bit rate and video quality level. This paper comprehensively reviews the classic RC technologies used in international video standards of past generations, analyses the mathematical models and implementation mechanisms of various schemes, and compares the performance of recent state-of-the-art RC algorithms. Finally, we discuss future directions and new application areas for RC methods. We hope that this review can help support the development, implementation, and application of RC for new video coding standards.

cs.MM↗