arXiv subjects

Chengcheng Han

Publications and source records attributed to Chengcheng Han.

At least 37 records · Page 2Linked to original sources

arXiv2025

Superhorizon fluctuations and the cosmic dipole problem

Recent observations have identified a significant 4.9$σ$ tension between the cosmic dipole inferred from galaxy number counts and that derived from the Cosmic Microwave Background (CMB), suggesting a potential deviation from the cosmological principle. This work investigates whether superhorizon isocurvature perturbations in cold dark matter (CDM) can account for this discrepancy. We demonstrate that, unlike adiabatic modes which cancel at leading order, superhorizon isocurvature modes can generate an intrinsic CMB dipole without significantly affecting galaxy number counts, thereby explaining the observed mismatch. We explore both single-mode and continuous-spectrum cases, focusing on two concrete models: a nearly scale-invariant power-law spectrum with a UV cutoff and axion-induced isocurvature perturbations. For the axion scenario, we show that if the radial mode evolves during inflation, the resulting perturbations can match the required amplitude while evading current CMB constraints. Our analysis constrains the self-coupling of associated potential for the axion to the range $10^{-9} < λ< 4 \times 10^{-9}$. These findings offer a viable solution to the dipole tension and may serve as indirect evidence for axion dark matter.

astro-ph.CO↗

arXiv2025

Hybrid Type-II and Type-III seesaw model for the muon $g-2$ anomaly

In this work we investigate the muon anomalous magnetic dipole moment $a_μ$ in a model that extends the Standard Model with a scalar triplet and a lepton triplet. Different from previous studies, we find that there is still viable parameter space in this model to explain the discrepancy $Δa_μ=a_μ(\mathrm{Exp})-a_μ(\mathrm{SM})$. While being consistent with the current data of neutrino mass, electroweak precision measurements and the perturbativity of couplings, our model can provide new physics contribution $a_μ^\textrm{NP}$ to cover the central region of $Δa_μ$ with new scalar and lepton mass as low as around TeV. This mass scale is allowed by the current collider searches for doubly charged scalars and the lepton triplet, and they can be tested at future high energy and/or high luminosity colliders.

hep-ph↗

arXiv2025

Topological defects as effective dynamical dark energy

In this work, we consider the possibility that the dynamical dark energy hinted at by recent DESI data may be mimicked by the effects of additional components in the universe, potentially arising from topological defects. We find that the data does not show a particular preference for the existence of cosmic strings. However, a domain wall contribution at the percent level can improve the fit, yielding a $Δχ^2= -1.72$ compared to the $Λ\rm{CDM}$ model. The improvement indicates that topological defects remain a viable and interesting extension to $Λ\rm{CDM}$, meriting further investigation with future cosmological data.

hep-ph↗

arXiv2025

RMoA: Optimizing Mixture-of-Agents through Diversity Maximization and Residual Compensation

Although multi-agent systems based on large language models show strong capabilities on multiple tasks, they are still limited by high computational overhead, information loss, and robustness. Inspired by ResNet's residual learning, we propose Residual Mixture-of-Agents (RMoA), integrating residual connections to optimize efficiency and reliability. To maximize information utilization from model responses while minimizing computational costs, we innovatively design an embedding-based diversity selection mechanism that greedily selects responses via vector similarity. Furthermore, to mitigate iterative information degradation, we introduce a Residual Extraction Agent to preserve cross-layer incremental information by capturing inter-layer response differences, coupled with a Residual Aggregation Agent for hierarchical information integration. Additionally, we propose an adaptive termination mechanism that dynamically halts processing based on residual convergence, further improving inference efficiency. RMoA achieves state-of-the-art performance on the benchmarks of across alignment, mathematical reasoning, code generation, and multitasking understanding, while significantly reducing computational overhead. Code is available at https://github.com/mindhunter01/RMoA.

cs.AI↗

arXiv2025

Probing Non-Thermal Gravitinos Through Large-Scale Structure Observations

We investigate the effects of non-thermally produced dark matter on large-scale structure formation, focusing on the gravitino dark matter. Our analysis shows that large-scale structure measurements from the SDSS LRG data offer a promising approach to probing non-thermally produced gravitinos. Specifically, if gravitinos resulting from neutralino decay constitute a fraction $F_\text{dec}=0.35~(0.1)$ of the dark matter component, the data exclude bino-like neutralino with a mass $m_{χ_1^0}\lesssim 25~(11)\text{ GeV}$ at $95\%\text{ C.L.}$. Furthermore, this constraint appears to be largely independent of the gravitino mass.

hep-ph↗

arXiv2025

A Survey of Neural Code Intelligence: Paradigms, Advances and Beyond

Neural Code Intelligence -- leveraging deep learning to understand, generate, and optimize code -- holds immense potential for transformative impacts on the whole society. Bridging the gap between Natural Language and Programming Language, this domain has drawn significant attention from researchers in both research communities over the past few years. This survey presents a systematic and chronological review of the advancements in code intelligence, encompassing over 50 representative models and their variants, more than 20 categories of tasks, and an extensive coverage of over 680 related works. We follow the historical progression to trace the paradigm shifts across different research phases (e.g., from modeling code with recurrent neural networks to the era of Large Language Models). Concurrently, we highlight the major technical transitions in models, tasks, and evaluations spanning through different stages. For applications, we also observe a co-evolving shift. It spans from initial endeavors to tackling specific scenarios, through exploring a diverse array of tasks during its rapid expansion, to currently focusing on tackling increasingly complex and varied real-world challenges. Building on our examination of the developmental trajectories, we further investigate the emerging synergies between code intelligence and broader machine intelligence, uncovering new cross-domain opportunities and illustrating the substantial influence of code intelligence across various domains. Finally, we delve into both the opportunities and challenges associated with this field, alongside elucidating our insights on the most promising research directions. An ongoing, dynamically updated project and resources associated with this survey have been released at https://github.com/QiushiSun/Awesome-Code-Intelligence.

cs.SE↗

arXiv2025

Probing Type II Seesaw Leptogenesis Through Lepton Flavor Violation

Lepton flavor violation (LFV) offers a powerful probe of physics beyond the Standard Model, particularly in models addressing neutrino masses and the baryon asymmetry of the universe. In this study, we investigate LFV processes within the framework of type II seesaw leptogenesis, where the Standard Model is extended by an $SU(2)_L$ triplet Higgs field. We focus on key LFV processes including $μ^+\to e^+γ$, $μ^+ \to e^+e^-e^+$, and $μ\rightarrow e$ conversion in nuclei, deriving stringent constraints on the parameter space from current experimental data. We scan the 3$σ$ range of neutrino oscillation parameters and identify the most conservative bounds consistent with existing measurements. Our results reveal that the MEG experiment currently provides the strongest constraints in the normal ordering (NO) scenario, while the SINDRUM experiment offers comparable sensitivity in the inverted ordering (IO) case. Future experiments, such as MEG II, Mu3e, Mu2e, and COMET, are predicted to significantly improve the sensitivity, testing larger regions of the parameter space. This work underscores the crucial role of LFV experiments in probing type II seesaw leptogenesis, providing an avenue to explore the connections between neutrino mass generation, baryogenesis, and inflation at experimentally accessible energy scales.

hep-ph↗

arXiv2025

Self-interacting dark matter implied by nano-Hertz gravitational waves

The self-interacting dark matter (SIDM) paradigm offers a potential solution to the small-scale structure problems faced by collision-less cold dark matter. This framework incorporates self-interactions among dark matter particles, typically mediated by a particle with a MeV-scale mass. Recent evidences of nano-Hertz gravitational waves from pulsar timing arrays (PTAs) such as NANOGrav, CPTA, EPTA, and PPTA suggest the occurrence of a first-order phase transition (FOPT) at a MeV-scale temperature. Considering the close proximity between these two scales, we propose that the mediator mass in the SIDM model originates from the spontaneous breaking of a $U(1)'$ symmetry, which is driven by the FOPT indicated by PTA data. Consequently, the alignment of these two scales is believed to be deeply connected by the same underlying physics. By extensively exploring the parameter space, remarkably, we find that the parameter space favored by SIDM just provides an explanation for the PTA data.

hep-ph↗

arXiv2024

Length Desensitization in Direct Preference Optimization

Direct Preference Optimization (DPO) is widely utilized in the Reinforcement Learning from Human Feedback (RLHF) phase to align Large Language Models (LLMs) with human preferences, thereby enhancing both their harmlessness and efficacy. However, it has been observed that DPO tends to over-optimize for verbosity, which can detrimentally affect both performance and user experience. In this paper, we conduct an in-depth theoretical analysis of DPO's optimization objective and reveal a strong correlation between its implicit reward and data length. This correlation misguides the optimization direction, resulting in length sensitivity during the DPO training and leading to verbosity. To address this issue, we propose a length-desensitization improvement method for DPO, termed LD-DPO. The proposed method aims to desensitize DPO to data length by decoupling explicit length preference, which is relatively insignificant, from the other implicit preferences, thereby enabling more effective learning of the intrinsic preferences. We utilized two settings (Base and Instruct) of Llama2-13B, Llama3-8B, and Qwen2-7B for experimental validation on various benchmarks including MT-Bench and AlpacaEval 2. The experimental results indicate that LD-DPO consistently outperforms DPO and other baseline methods, achieving more concise responses with a 10-40% reduction in length compared to DPO. We conducted in-depth experimental analyses to demonstrate that LD-DPO can indeed achieve length desensitization and align the model more closely with human-like preferences.

cs.LG↗

arXiv2024

Automated Peer Reviewing in Paper SEA: Standardization, Evaluation, and Analysis

In recent years, the rapid increase in scientific papers has overwhelmed traditional review mechanisms, resulting in varying quality of publications. Although existing methods have explored the capabilities of Large Language Models (LLMs) for automated scientific reviewing, their generated contents are often generic or partial. To address the issues above, we introduce an automated paper reviewing framework SEA. It comprises of three modules: Standardization, Evaluation, and Analysis, which are represented by models SEA-S, SEA-E, and SEA-A, respectively. Initially, SEA-S distills data standardization capabilities of GPT-4 for integrating multiple reviews for a paper. Then, SEA-E utilizes standardized data for fine-tuning, enabling it to generate constructive reviews. Finally, SEA-A introduces a new evaluation metric called mismatch score to assess the consistency between paper contents and reviews. Moreover, we design a self-correction strategy to enhance the consistency. Extensive experimental results on datasets collected from eight venues show that SEA can generate valuable insights for authors to improve their papers.

cs.CL↗

arXiv2024

Learning to Check: Unleashing Potentials for Self-Correction in Large Language Models

Self-correction has achieved impressive results in enhancing the style and security of the generated output from large language models (LLMs). However, recent studies suggest that self-correction might be limited or even counterproductive in reasoning tasks due to LLMs' difficulties in identifying logical mistakes. In this paper, we aim to enhance the self-checking capabilities of LLMs by constructing training data for checking tasks. Specifically, we apply the Chain of Thought (CoT) methodology to self-checking tasks, utilizing fine-grained step-level analyses and explanations to assess the correctness of reasoning paths. We propose a specialized checking format called "Step CoT Check". Following this format, we construct a checking-correction dataset that includes detailed step-by-step analysis and checking. Then we fine-tune LLMs to enhance their error detection and correction abilities. Our experiments demonstrate that fine-tuning with the "Step CoT Check" format significantly improves the self-checking and self-correction abilities of LLMs across multiple benchmarks. This approach outperforms other formats, especially in locating the incorrect position, with greater benefits observed in larger models. For reproducibility, all the datasets and code are provided in https://github.com/bammt/Learn-to-check.

cs.CL↗

arXiv2024

Make Prompt-based Black-Box Tuning Colorful: Boosting Model Generalization from Three Orthogonal Perspectives

Large language models (LLMs) have shown increasing power on various natural language processing (NLP) tasks. However, tuning these models for downstream tasks usually needs exorbitant costs or is unavailable due to commercial considerations. Recently, black-box tuning has been proposed to address this problem by optimizing task-specific prompts without accessing the gradients and hidden representations. However, most existing works have yet fully exploited the potential of gradient-free optimization under the scenario of few-shot learning. In this paper, we describe BBT-RGB, a suite of straightforward and complementary techniques for enhancing the efficiency and performance of black-box optimization. Specifically, our method includes three plug-and-play components: (1) Two-stage derivative-free optimization strategy that facilitates fast convergence and mitigates overfitting; (2) Automatic verbalizer construction with its novel usage under few-shot settings; (3) Better prompt initialization policy based on instruction search and auto-selected demonstration. Extensive experiments across various tasks on natural language understanding and inference demonstrate the effectiveness of our method. Our codes are publicly available at https://github.com/QiushiSun/BBT-RGB.

cs.CL↗

arXiv2024

Affleck-Dine Dirac Leptogenesis

We present a minimal framework that realises successful Dirac Leptogenesis through the Affleck-Dine mechanism. A single right-handed neutrino and a neutrinophillic Higgs doublet are introduced to the Standard Model, which couple via a Yukawa interaction. The inflationary setting is induced by a combination of the two Higgs doublets, with their global symmetry violating interactions leading to a net charge generation via the Affleck-Dine mechanism. This simple Standard Model extension exhibits a unique and connected set of phenomenological implications including the resultant baryon asymmetry, inflationary predictions, cosmological implications, relic right-handed neutrinos, and its low energy phenomenology, while also being able to be embedded in various neutrino mass generating mechanisms.

hep-ph↗

arXiv2024

OS-Copilot: Towards Generalist Computer Agents with Self-Improvement

Autonomous interaction with the computer has been a longstanding challenge with great potential, and the recent proliferation of large language models (LLMs) has markedly accelerated progress in building digital agents. However, most of these agents are designed to interact with a narrow domain, such as a specific software or website. This narrow focus constrains their applicability for general computer tasks. To this end, we introduce OS-Copilot, a framework to build generalist agents capable of interfacing with comprehensive elements in an operating system (OS), including the web, code terminals, files, multimedia, and various third-party applications. We use OS-Copilot to create FRIDAY, a self-improving embodied agent for automating general computer tasks. On GAIA, a general AI assistants benchmark, FRIDAY outperforms previous methods by 35%, showcasing strong generalization to unseen applications via accumulated skills from previous tasks. We also present numerical and quantitative evidence that FRIDAY learns to control and self-improve on Excel and Powerpoint with minimal supervision. Our OS-Copilot framework and empirical findings provide infrastructure and insights for future research toward more capable and general-purpose computer agents.

cs.AI↗

arXiv2023

Freeze-in bino dark matter in high scale supersymmetry

We explore a scenario of high scale supersymmetry where all supersymmetric particles except gauginos stay at a high energy scale $M_{\rm SUSY}$ which is much larger than the reheating temperature $T_\text{RH}$. The dark matter is dominated by bino component with mass around the electroweak scale and the observed relic abundance is mainly generated by the freeze-in process during the early universe. Considering the various constraints, we identify two available scenarios in which the supersymmetric sector at an energy scale below $T_\text{RH}$ consists of: a) bino; b) bino and wino. Typically, for a bino mass around 0.1-1 TeV and a wino mass around 2 TeV, we find that $M_{\rm SUSY}$ should be around $10^{12-14}$ GeV with $T_\text{RH}$ around $10^{4-6}$ GeV.

hep-ph↗

arXiv2023

Type-II Seesaw Leptogenesis along the Ridge

Type-II seesaw leptogenesis is a model that integrates inflation, baryon number asymmetry, and neutrino mass simultaneously. It employs the Affleck-Dine mechanism to generate lepton asymmetry, with the Higgs bosons serving as the inflaton. Previous studies assumed inflation to occur in a valley of the potential, employing the single-field approximation. In this work, we explore an alternative scenario for the type-II seesaw leptogenesis, where the inflation takes place along a ridge of the potential. Firstly, we conduct a comprehensive numerical calculation in the canonical scenario, where inflation occurs in a valley, confirming the effectiveness of the single-field approximation. Then, we introduce a novel scenario wherein inflation initiates along the potential's ridge and transitions to the valley in the late stages. In this case, the single-field inflation approximation is no longer valid, yet leptogenesis is still successfully achieved. We find that this scenario can generate a significant non-Gaussianity signature, offering testable predictions for future experiments.

hep-ph↗

arXiv2023

DialCoT Meets PPO: Decomposing and Exploring Reasoning Paths in Smaller Language Models

Chain-of-Thought (CoT) prompting has proven to be effective in enhancing the reasoning capabilities of Large Language Models (LLMs) with at least 100 billion parameters. However, it is ineffective or even detrimental when applied to reasoning tasks in Smaller Language Models (SLMs) with less than 10 billion parameters. To address this limitation, we introduce Dialogue-guided Chain-of-Thought (DialCoT) which employs a dialogue format to generate intermediate reasoning steps, guiding the model toward the final answer. Additionally, we optimize the model's reasoning path selection using the Proximal Policy Optimization (PPO) algorithm, further enhancing its reasoning capabilities. Our method offers several advantages compared to previous approaches. Firstly, we transform the process of solving complex reasoning questions by breaking them down into a series of simpler sub-questions, significantly reducing the task difficulty and making it more suitable for SLMs. Secondly, we optimize the model's reasoning path selection through the PPO algorithm. We conduct comprehensive experiments on four arithmetic reasoning datasets, demonstrating that our method achieves significant performance improvements compared to state-of-the-art competitors.

cs.CL↗

arXiv2023

Exchanging-based Multimodal Fusion with Transformer

We study the problem of multimodal fusion in this paper. Recent exchanging-based methods have been proposed for vision-vision fusion, which aim to exchange embeddings learned from one modality to the other. However, most of them project inputs of multimodalities into different low-dimensional spaces and cannot be applied to the sequential input data. To solve these issues, in this paper, we propose a novel exchanging-based multimodal fusion model MuSE for text-vision fusion based on Transformer. We first use two encoders to separately map multimodal inputs into different low-dimensional spaces. Then we employ two decoders to regularize the embeddings and pull them into the same space. The two decoders capture the correlations between texts and images with the image captioning task and the text-to-image generation task, respectively. Further, based on the regularized embeddings, we present CrossTransformer, which uses two Transformer encoders with shared parameters as the backbone model to exchange knowledge between multimodalities. Specifically, CrossTransformer first learns the global contextual information of the inputs in the shallow layers. After that, it performs inter-modal exchange by selecting a proportion of tokens in one modality and replacing their embeddings with the average of embeddings in the other modality. We conduct extensive experiments to evaluate the performance of MuSE on the Multimodal Named Entity Recognition task and the Multimodal Sentiment Analysis task. Our results show the superiority of MuSE against other competitors. Our code and data are provided at https://github.com/RecklessRonan/MuSE.

cs.CV↗