arXiv subjects

Xinyu Dai

Publications and source records attributed to Xinyu Dai.

At least 55 records · Page 3Linked to original sources

arXiv2025

Comparing Optical Variability of Type 1 and Type 2 AGN from the BAT 9-Month Sample using ASAS-SN and TESS Surveys

We present an optical variability analysis and comparison of the samples of Seyfert 1 and 2 galaxies, selected from the \textit{Swift} 9-month BAT catalog, using the light curves from Transiting Exoplanet Survey Satellite (TESS) and All-Sky Automated Survey for SuperNovae (ASAS-SN). We measured the normalized excess variance of TESS and ASAS-SN light curves for each target and performed a Kolmogorov-Smirnov test between the two samples, where our results showed significant differences. This is consistent with predictions from the unification model, where Seyfert 2s are obscured by the larger scale dust torus and their variability is suppressed. This variability difference is independent of the luminosity, Eddington ratio, or black hole mass, further supporting geometrical unification models. We searched the dependence of the normalized excess variance of Seyfert 1s on absolute magnitudes, Eddington ratio, and black hole mass, where our results are consistent with relations found in the literature. Finally, a small sub-sample of changing-look AGNs that transitioned during the time frame of the ASAS-SN light curves, with their variability amplitudes changing according to the classification, larger variability as type 1s and smaller as 2s. The change of variability amplitudes can be used to better pinpoint when the type transition occurred. The consistency trend of the variability amplitude differences between Seyfert 1s and 2s and between changing-look AGNs in 1 or 2 stages suggests that variability can be a key factor in shedding light on the changing-look AGN or the dichotomy between Seyfert 1 or 2 populations.

astro-ph.GA↗

arXiv2025

High frequency breaks in optical active galactic nuclei power spectral density

Context. Variability is a ubiquitous feature of active galactic nuclei (AGNs), and the characterisation of the variability is crucial to constrain its physical mechanism and proper applications in AGN studies. The advent of all-sky and high-cadence optical surveys allows more accurate measurements of AGNs variability down to short timescales and direct comparisons with X-ray variability from the same sample of sources. Aims. We aim to analyse the optical power spectral density (PSDs) of AGNs with measured X-ray PSDs. Methods. We use light curves from the All-Sky Automated Survey for SuperNovae (ASAS-SN) and the Transiting Exoplanet Survey Satellite (TESS) and use the Lomb-Scargle periodogram to obtain PSDs. The joint optical PSD is measured over up to six orders of magnitude in frequency space from timescales of minutes to a decade. We fit either a damped random walk (DRW) or a broken power law (BPL) model to constrain the PSD model and break frequency. Results. We find a set of break frequencies ($\lesssim10^{-2}$ day$^{-1}$) from DRW and BPL fits that generally confirm previously reported correlations between break frequencies and the black hole mass. In addition, we find a second set of break frequencies at higher frequencies ($>10^{-2}$ day$^{-1}$). We observe a potential weak correlation between the high-frequency breaks with the X-ray break frequencies and black hole mass. We further explore the dependence of the correlations on other AGN parameters, finding that adding either X-ray, optical, or bolometric luminosity as the third correlation parameter can substantially improve the correlation significances. The newly identified high-frequency optical breaks can constrain different aspects of the physics of active galactic nuclei.

astro-ph.HE↗

arXiv2025

Multi-wavelength and Environmental Properties of Variability Selected Low Luminosity Active Galactic Nuclei

We present the multi-wavelength and environmental properties of 37 variability-selected active galactic nuclei (AGNs), including 30 low luminosity AGNs (LLAGNs), using a high cadence time-domain survey (ASAS-SN) from a spectroscopic sample of 1218 nearby bright galaxies. We find that high-cadence time-domain surveys uniquely select LLAGNs that do not necessarily satisfy other AGN selection methods, such as X-ray, mid-IR, or BPT methods. In our sample, 3% of them pass the mid-infrared color based AGN selection, 18% pass the X-ray luminosity based AGN selection, and 60% pass the BPT selection. This result is supported by two other LLAGN samples from high-cadence time-domain surveys of TESS and PTF, suggesting that the variability selection method from well-sampled light curves can find AGNs that may not be discovered otherwise. These AGNs can have moderate to small amplitudes of variability from the accretion disk, but, of many of them, with no strong corona, emission lines from the central engine, or accretion power to dominate the mid-IR emission. The X-ray spectra of a sub-sample of bright sources are consistent with a power law model. Upon inspecting the environments of our sample, we find that LLAGNs are more common in denser environments of galaxy clusters in contrast with the trend established in the literature for luminous AGNs at low redshifts, which is broadly consistent with our analysis result for luminous AGNs limited by a smaller sample size. This contrast in environmental properties between LLAGN and luminous AGNs suggests that LLAGNs may have different trigger mechanisms.

astro-ph.GA↗

arXiv2025

The AGN Optical Variability Fundamental Plane

We investigate the relationship between AGN optical variability timescales, amplitudes, and supermassive black hole (SMBH) masses using homogeneous light curves from the All-Sky Automated Survey for SuperNovae (ASAS-SN). We fit a damped random walk (DRW) model to high-cadence, long-baseline ASAS-SN light curves to estimate the characteristic variability timescale ($τ_\text{DRW}$) and amplitude ($σ$) for 57 AGN with precise SMBH mass measurements from reverberation mapping and dynamical methods. We confirm a significant correlation between $τ_\text{DRW}$ and SMBH mass, and find: $\text{log}_{10}(M_\text{BH}/ \text{M}_\odot) = (1.85\pm0.20)\times\text{log}_{10} (τ_\text{DRW}/200 \text{ days})+7.59\pm0.08$. Incorporating $\hatσ^2 = 2σ^2/τ_\text{DRW}$ in a plane model significantly improves residuals, and we find: $\text{log}_{10}(M_\text{BH}/ \text{M}_\odot) = (2.27\pm0.20)\times\text{log}_{10} (τ_\text{DRW}/200\text{ days})+(1.20\pm0.20)\times\text{log}_{10}(\hatσ/\text{1 mJy/days}^{1/2})+7.68\pm0.08$ with a scatter of 0.39 dex. We calculate $τ_\text{DRW}$, $\hatσ$, and estimate SMBH masses for 203 bright ($V<16$ mag) AGN from the Milliquas catalog and compare these estimates with measurements from the BAT AGN Spectroscopic Survey for 42 overlapping AGN. In 10 years, LSST could extend this method to survey $7\lesssim\text{log}_{10}({M_\text{BH}/M_\odot})\lesssim9$ SMBHs out to $z\sim1$ and $\textrm{log}_{10}({M_\text{BH}/M_\odot})\sim8.0$ out to $z\sim4$, and ASAS-SN could probe $5\lesssim \textrm{log}_{10}({M_\text{BH}/M_\odot})\lesssim10.5$ SMBHs in the local universe and $\textrm{log}_{10}({M_\text{BH}/M_\odot})\sim9.0$ out to $z\sim2$. Measuring AGN variability with these datasets will provide a unique probe of SMBH evolution by making estimates of $M_\text{BH}$ spanning several orders of magnitude with photometric observations alone.

astro-ph.GA↗

arXiv2025

Rethinking Relation Extraction: Beyond Shortcuts to Generalization with a Debiased Benchmark

Benchmarks are crucial for evaluating machine learning algorithm performance, facilitating comparison and identifying superior solutions. However, biases within datasets can lead models to learn shortcut patterns, resulting in inaccurate assessments and hindering real-world applicability. This paper addresses the issue of entity bias in relation extraction tasks, where models tend to rely on entity mentions rather than context. We propose a debiased relation extraction benchmark DREB that breaks the pseudo-correlation between entity mentions and relation types through entity replacement. DREB utilizes Bias Evaluator and PPL Evaluator to ensure low bias and high naturalness, providing a reliable and accurate assessment of model generalization in entity bias scenarios. To establish a new baseline on DREB, we introduce MixDebias, a debiasing method combining data-level and model training-level techniques. MixDebias effectively improves model performance on DREB while maintaining performance on the original dataset. Extensive experiments demonstrate the effectiveness and robustness of MixDebias compared to existing methods, highlighting its potential for improving the generalization ability of relation extraction models. We will release DREB and MixDebias publicly.

cs.AI↗

arXiv2024

Dynamic Demonstrations Controller for In-Context Learning

In-context learning (ICL) is a new paradigm for natural language processing (NLP), where a large language model (LLM) observes a small number of demonstrations and a test instance as its input, and directly makes predictions without updating model parameters. Previous studies have revealed that ICL is sensitive to the selection and the ordering of demonstrations. However, there are few studies regarding the impact of the demonstration number on the ICL performance within a limited input length of LLM, because it is commonly believed that the number of demonstrations is positively correlated with model performance. In this paper, we found this conclusion does not always hold true. Through pilot experiments, we discover that increasing the number of demonstrations does not necessarily lead to improved performance. Building upon this insight, we propose a Dynamic Demonstrations Controller (D$^2$Controller), which can improve the ICL performance by adjusting the number of demonstrations dynamically. The experimental results show that D$^2$Controller yields a 4.6% relative improvement on ten different sizes of LLMs across ten datasets. Moreover, we also extend our method to previous ICL models and achieve competitive results.

cs.CL↗

arXiv2024

AlignGPT: Multi-modal Large Language Models with Adaptive Alignment Capability

Multimodal Large Language Models (MLLMs) are widely regarded as crucial in the exploration of Artificial General Intelligence (AGI). The core of MLLMs lies in their capability to achieve cross-modal alignment. To attain this goal, current MLLMs typically follow a two-phase training paradigm: the pre-training phase and the instruction-tuning phase. Despite their success, there are shortcomings in the modeling of alignment capabilities within these models. Firstly, during the pre-training phase, the model usually assumes that all image-text pairs are uniformly aligned, but in fact the degree of alignment between different image-text pairs is inconsistent. Secondly, the instructions currently used for finetuning incorporate a variety of tasks and different tasks usually require different levels of alignment capabilities, but previous MLLMs overlook these differentiated alignment needs. To tackle these issues, we propose a new multimodal large language model AlignGPT. In the pre-training stage, instead of treating all image-text pairs equally, we divide them into different groups according to the degrees of alignment of them. Then, the model is trained to learn the representations of different alignment levels. In the instruction-tuning phase, we adaptively combine these representations of alignment levels to meet the dynamic alignment needs of different tasks. Extensive experimental results show that our model achieves competitive performance on 12 benchmarks.

cs.CL↗

arXiv2024

PreAlign: Boosting Cross-Lingual Transfer by Early Establishment of Multilingual Alignment

Large language models demonstrate reasonable multilingual abilities, despite predominantly English-centric pretraining. However, the spontaneous multilingual alignment in these models is shown to be weak, leading to unsatisfactory cross-lingual transfer and knowledge sharing. Previous works attempt to address this issue by explicitly injecting multilingual alignment information during or after pretraining. Thus for the early stage in pretraining, the alignment is weak for sharing information or knowledge across languages. In this paper, we propose PreAlign, a framework that establishes multilingual alignment prior to language model pretraining. PreAlign injects multilingual alignment by initializing the model to generate similar representations of aligned words and preserves this alignment using a code-switching strategy during pretraining. Extensive experiments in a synthetic English to English-Clone setting demonstrate that PreAlign significantly outperforms standard multilingual joint training in language modeling, zero-shot cross-lingual transfer, and cross-lingual knowledge application. Further experiments in real-world scenarios further validate PreAlign's effectiveness across various model sizes.

cs.CL↗

arXiv2024

EFUF: Efficient Fine-grained Unlearning Framework for Mitigating Hallucinations in Multimodal Large Language Models

Multimodal large language models (MLLMs) have attracted increasing attention in the past few years, but they may still generate descriptions that include objects not present in the corresponding images, a phenomenon known as object hallucination. To eliminate hallucinations, existing methods manually annotate paired responses with and without hallucinations, and then employ various alignment algorithms to improve the alignment capability between images and text. However, they not only demand considerable computation resources during the finetuning stage but also require expensive human annotation to construct paired data needed by the alignment algorithms. To address these issues, we borrow the idea of unlearning and propose an efficient fine-grained unlearning framework (EFUF), which can eliminate hallucinations without the need for paired data. Extensive experiments show that our method consistently reduces hallucinations while preserving the generation quality with modest computational overhead. Our code and datasets will be publicly available.

cs.CL↗

arXiv2024

RAGLAB: A Modular and Research-Oriented Unified Framework for Retrieval-Augmented Generation

Large Language Models (LLMs) demonstrate human-level capabilities in dialogue, reasoning, and knowledge retention. However, even the most advanced LLMs face challenges such as hallucinations and real-time updating of their knowledge. Current research addresses this bottleneck by equipping LLMs with external knowledge, a technique known as Retrieval Augmented Generation (RAG). However, two key issues constrained the development of RAG. First, there is a growing lack of comprehensive and fair comparisons between novel RAG algorithms. Second, open-source tools such as LlamaIndex and LangChain employ high-level abstractions, which results in a lack of transparency and limits the ability to develop novel algorithms and evaluation metrics. To close this gap, we introduce RAGLAB, a modular and research-oriented open-source library. RAGLAB reproduces 6 existing algorithms and provides a comprehensive ecosystem for investigating RAG algorithms. Leveraging RAGLAB, we conduct a fair comparison of 6 RAG algorithms across 10 benchmarks. With RAGLAB, researchers can efficiently compare the performance of various algorithms and develop novel algorithms.

cs.CL↗

arXiv2024

AutoSurvey: Large Language Models Can Automatically Write Surveys

This paper introduces AutoSurvey, a speedy and well-organized methodology for automating the creation of comprehensive literature surveys in rapidly evolving fields like artificial intelligence. Traditional survey paper creation faces challenges due to the vast volume and complexity of information, prompting the need for efficient survey methods. While large language models (LLMs) offer promise in automating this process, challenges such as context window limitations, parametric knowledge constraints, and the lack of evaluation benchmarks remain. AutoSurvey addresses these challenges through a systematic approach that involves initial retrieval and outline generation, subsection drafting by specialized LLMs, integration and refinement, and rigorous evaluation and iteration. Our contributions include a comprehensive solution to the survey problem, a reliable evaluation method, and experimental validation demonstrating AutoSurvey's effectiveness.We open our resources at \url{https://github.com/AutoSurveys/AutoSurvey}.

cs.IR↗

arXiv2024

ESM All-Atom: Multi-scale Protein Language Model for Unified Molecular Modeling

Protein language models have demonstrated significant potential in the field of protein engineering. However, current protein language models primarily operate at the residue scale, which limits their ability to provide information at the atom level. This limitation prevents us from fully exploiting the capabilities of protein language models for applications involving both proteins and small molecules. In this paper, we propose ESM-AA (ESM All-Atom), a novel approach that enables atom-scale and residue-scale unified molecular modeling. ESM-AA achieves this by pre-training on multi-scale code-switch protein sequences and utilizing a multi-scale position encoding to capture relationships among residues and atoms. Experimental results indicate that ESM-AA surpasses previous methods in protein-molecule tasks, demonstrating the full utilization of protein language models. Further investigations reveal that through unified molecular modeling, ESM-AA not only gains molecular knowledge but also retains its understanding of proteins. The source codes of ESM-AA are publicly released at https://github.com/zhengkangjie/ESM-AA.

q-bio.BM↗

arXiv2024

Extroversion or Introversion? Controlling The Personality of Your Large Language Models

Large language models (LLMs) exhibit robust capabilities in text generation and comprehension, mimicking human behavior and exhibiting synthetic personalities. However, some LLMs have displayed offensive personality, propagating toxic discourse. Existing literature neglects the origin and evolution of LLM personalities, as well as the effective personality control. To fill these gaps, our study embarked on a comprehensive investigation into LLM personality control. We investigated several typical methods to influence LLMs, including three training methods: Continual Pre-training, Supervised Fine-Tuning (SFT), and Reinforcement Learning from Human Feedback (RLHF), along with inference phase considerations (prompts). Our investigation revealed a hierarchy of effectiveness in control: Prompt > SFT > RLHF > Continual Pre-train. Notably, SFT exhibits a higher control success rate compared to prompt induction. While prompts prove highly effective, we found that prompt-induced personalities are less robust than those trained, making them more prone to showing conflicting personalities under reverse personality prompt induction. Besides, harnessing the strengths of both SFT and prompt, we proposed $\underline{\text{P}}$rompt $\underline{\text{I}}$nduction post $\underline{\text{S}}$upervised $\underline{\text{F}}$ine-tuning (PISF), which emerges as the most effective and robust strategy for controlling LLMs' personality, displaying high efficacy, high success rates, and high robustness. Even under reverse personality prompt induction, LLMs controlled by PISF still exhibit stable and robust personalities.

cs.CL↗

arXiv2024

The Devil is in the Few Shots: Iterative Visual Knowledge Completion for Few-shot Learning

Contrastive Language-Image Pre-training (CLIP) has shown powerful zero-shot learning performance. Few-shot learning aims to further enhance the transfer capability of CLIP by giving few images in each class, aka 'few shots'. Most existing methods either implicitly learn from the few shots by incorporating learnable prompts or adapters, or explicitly embed them in a cache model for inference. However, the narrow distribution of few shots often contains incomplete class information, leading to biased visual knowledge with high risk of misclassification. To tackle this problem, recent methods propose to supplement visual knowledge by generative models or extra databases, which can be costly and time-consuming. In this paper, we propose an Iterative Visual Knowledge CompLetion (KCL) method to complement visual knowledge by properly taking advantages of unlabeled samples without access to any auxiliary or synthetic data. Specifically, KCL first measures the similarities between unlabeled samples and each category. Then, the samples with top confidence to each category is selected and collected by a designed confidence criterion. Finally, the collected samples are treated as labeled ones and added to few shots to jointly re-estimate the remaining unlabeled ones. The above procedures will be repeated for a certain number of iterations with more and more samples being collected until convergence, ensuring a progressive and robust knowledge completion process. Extensive experiments on 11 benchmark datasets demonstrate the effectiveness and efficiency of KCL as a plug-and-play module under both few-shot and zero-shot learning settings. Code is available at https://github.com/Mark-Sky/KCL.

cs.CV↗

arXiv2024

Emotion-Anchored Contrastive Learning Framework for Emotion Recognition in Conversation

Emotion Recognition in Conversation (ERC) involves detecting the underlying emotion behind each utterance within a conversation. Effectively generating representations for utterances remains a significant challenge in this task. Recent works propose various models to address this issue, but they still struggle with differentiating similar emotions such as excitement and happiness. To alleviate this problem, We propose an Emotion-Anchored Contrastive Learning (EACL) framework that can generate more distinguishable utterance representations for similar emotions. To achieve this, we utilize label encodings as anchors to guide the learning of utterance representations and design an auxiliary loss to ensure the effective separation of anchors for similar emotions. Moreover, an additional adaptation process is proposed to adapt anchors to serve as effective classifiers to improve classification performance. Across extensive experiments, our proposed EACL achieves state-of-the-art emotion recognition performance and exhibits superior performance on similar emotions. Our code is available at https://github.com/Yu-Fangxu/EACL.

cs.CL↗

arXiv2024

Knowledge-aware Dual-side Attribute-enhanced Recommendation

\textit{Knowledge-aware} recommendation methods (KGR) based on \textit{graph neural networks} (GNNs) and \textit{contrastive learning} (CL) have achieved promising performance. However, they fall short in modeling fine-grained user preferences and further fail to leverage the \textit{preference-attribute connection} to make predictions, leading to sub-optimal performance. To address the issue, we propose a method named \textit{\textbf{K}nowledge-aware \textbf{D}ual-side \textbf{A}ttribute-enhanced \textbf{R}ecommendation} (KDAR). Specifically, we build \textit{user preference representations} and \textit{attribute fusion representations} upon the attribute information in knowledge graphs, which are utilized to enhance \textit{collaborative filtering} (CF) based user and item representations, respectively. To discriminate the contribution of each attribute in these two types of attribute-based representations, a \textit{multi-level collaborative alignment contrasting} mechanism is proposed to align the importance of attributes with CF signals. Experimental results on four benchmark datasets demonstrate the superiority of KDAR over several state-of-the-art baselines. Further analyses verify the effectiveness of our method. The code of KDAR is released at: \href{https://github.com/TJTP/KDAR}{https://github.com/TJTP/KDAR}.

cs.IR↗

arXiv2024

Dr3: Ask Large Language Models Not to Give Off-Topic Answers in Open Domain Multi-Hop Question Answering

Open Domain Multi-Hop Question Answering (ODMHQA) plays a crucial role in Natural Language Processing (NLP) by aiming to answer complex questions through multi-step reasoning over retrieved information from external knowledge sources. Recently, Large Language Models (LLMs) have demonstrated remarkable performance in solving ODMHQA owing to their capabilities including planning, reasoning, and utilizing tools. However, LLMs may generate off-topic answers when attempting to solve ODMHQA, namely the generated answers are irrelevant to the original questions. This issue of off-topic answers accounts for approximately one-third of incorrect answers, yet remains underexplored despite its significance. To alleviate this issue, we propose the Discriminate->Re-Compose->Re- Solve->Re-Decompose (Dr3) mechanism. Specifically, the Discriminator leverages the intrinsic capabilities of LLMs to judge whether the generated answers are off-topic. In cases where an off-topic answer is detected, the Corrector performs step-wise revisions along the reversed reasoning chain (Re-Compose->Re-Solve->Re-Decompose) until the final answer becomes on-topic. Experimental results on the HotpotQA and 2WikiMultiHopQA datasets demonstrate that our Dr3 mechanism considerably reduces the occurrence of off-topic answers in ODMHQA by nearly 13%, improving the performance in Exact Match (EM) by nearly 3% compared to the baseline method without the Dr3 mechanism.

cs.CL↗

arXiv2024

Multi-Candidate Speculative Decoding

Large language models have shown impressive capabilities across a variety of NLP tasks, yet their generating text autoregressively is time-consuming. One way to speed them up is speculative decoding, which generates candidate segments (a sequence of tokens) from a fast draft model that is then verified in parallel by the target model. However, the acceptance rate of candidate tokens receives limitations from several factors, such as the model, the dataset, and the decoding setup. This paper proposes sampling multiple candidates from a draft model and then organising them in batches for verification. We design algorithms for efficient multi-candidate verification while maintaining the distribution of the target model. Our approach shows significant improvements in acceptance rates on multiple datasets and models, consistently outperforming standard speculative decoding.

cs.CL↗