arXiv subjects

Yi Cai

Publications and source records attributed to Yi Cai.

At least 73 records · Page 4Linked to original sources

arXiv2024

Learning to Correction: Explainable Feedback Generation for Visual Commonsense Reasoning Distractor

Large multimodal models (LMMs) have shown remarkable performance in the visual commonsense reasoning (VCR) task, which aims to answer a multiple-choice question based on visual commonsense within an image. However, the ability of LMMs to correct potential visual commonsense errors in the distractor upon their occurrence is yet under-explored. Drawing inspiration from how a human teacher crafts challenging distractors to test students' comprehension of the concepts or skills and assists them in identifying and correcting errors toward the answer, we are the pioneering research for LMMs to simulate this error correction process. To this end, we employ GPT-4 as a ``teacher'' to collect the explainable feedback dataset VCR-DF for error correction, which serves as a benchmark to evaluate the ability of LMMs to identify misconceptions and clarify reasons behind the error in VCR distractors toward final answers. In addition, we propose an LMM-based Pedagogical Expert Instructed Feedback Generation (PEIFG) model to incorporate the learnable expert prompts and multimodal instruction as guidance for feedback generation. Experimental results show that our PEIFG significantly outperforms existing LMMs. We believe that our benchmark provides a new direction for evaluating the capabilities of LMMs.

cs.CV↗

arXiv2024

Transparent Neighborhood Approximation for Text Classifier Explanation

Recent literature highlights the critical role of neighborhood construction in deriving model-agnostic explanations, with a growing trend toward deploying generative models to improve synthetic instance quality, especially for explaining text classifiers. These approaches overcome the challenges in neighborhood construction posed by the unstructured nature of texts, thereby improving the quality of explanations. However, the deployed generators are usually implemented via neural networks and lack inherent explainability, sparking arguments over the transparency of the explanation process itself. To address this limitation while preserving neighborhood quality, this paper introduces a probability-based editing method as an alternative to black-box text generators. This approach generates neighboring texts by implementing manipulations based on in-text contexts. Substituting the generator-based construction process with recursive probability-based editing, the resultant explanation method, XPROB (explainer with probability-based editing), exhibits competitive performance according to the evaluation conducted on two real-world datasets. Additionally, XPROB's fully transparent and more controllable construction process leads to superior stability compared to the generator-based explainers.

cs.CL↗

arXiv2024

Disco Intelligent Omni-Surfaces: 360-degree Fully-Passive Jamming Attacks

Intelligent omni-surfaces (IOSs) with 360-degree electromagnetic radiation significantly improves the performance of wireless systems, while an adversarial IOS also poses a significant potential risk for physical layer security. In this paper, we propose a "DISCO" IOS (DIOS) based fully-passive jammer (FPJ) that can launch omnidirectional fully-passive jamming attacks. In the proposed DIOS-based FPJ, the interrelated refractive and reflective (R&R) coefficients of the adversarial IOS are randomly generated, acting like a "DISCO" that distributes wireless energy radiated by the base station. By introducing active channel aging (ACA) during channel coherence time, the DIOS-based FPJ can perform omnidirectional fully-passive jamming without neither jamming power nor channel knowledge of legitimate users (LUs). To characterize the impact of the DIOS-based PFJ, we derive the statistical characteristics of DIOS-jammed channels based on two widely-used IOS models, i.e., the constant-amplitude model and the variable-amplitude model. Consequently, the asymptotic analysis of the ergodic achievable sum rates under the DIOS-based omnidirectional fully-passive jamming is given based on the derived stochastic characteristics for both the two IOS models. Based on the derived analysis, the omnidirectional jamming impact of the proposed DIOS-based FPJ implemented by a constant-amplitude IOS does not depend on either the quantization number or the stochastic distribution of the DIOS coefficients, while the conclusion does not hold on when a variable-amplitude IOS is used. Numerical results based on one-bit quantization of the IOS phase shifts are provided to verify the effectiveness of the derived theoretical analysis. The proposed DIOS-based FPJ can not only launch omnidirectional fully-passive jamming, but also improve the jamming impact by about 55% at 10 dBm transmit power per LU.

eess.SP↗

arXiv2024

Unlocking the Power of GANs in Non-Autoregressive Text Generation

Generative Adversarial Networks (GANs) have been studied in text generation to tackle the exposure bias problem. Despite their remarkable development, they adopt autoregressive structures so suffering from high latency in both training and inference stages. Although GANs have potential to support efficient generation by adopting non-autoregressive (NAR) structures, their explorations in NAR models are extremely limited. In this work, we conduct pioneering study of building language GANs based on NAR structures. We identify two issues that constrain the performance of GAN-based NAR models. Firstly, existing methods of incorporating latent variables provide highly similar representations which cannot describe the diversity of different words in sentences. We tackle this problem by proposing Position-Aware Self-Modulation, providing more diverse and effective representations. Secondly, the attention mechanism in Transformer cannot accurately build word dependencies in the unstable training of GANs, and we adopt Dependency Feed Forward Network to enhance the model capacity in dependency modeling. Armed with these two facilities, we propose a GAN-based NAR model, Adversarial Non-autoregressive Transformer (ANT). The experimental results demonstrate that ANT can achieve comparable performance with mainstream models in a single forward pass and has great potential in various applications like latent interpolation and semi-supervised learning.

cs.CL↗

arXiv2024

Lunar Swirls Unveil the Origin of the Moon Magnetic Field

The origins of the lunar magnetic anomalies and swirls have long puzzled scientists.The prevailing theory posits that an ancient lunar dynamo core field magnetized extralunar meteoritic materials, leading to the current remnant magnetic anomalies that shield against solar wind ions, thereby contributing to the formation of lunar swirls. Our research reveals that these lunar swirls are the result of ancient electrical currents that traversed the Moon's surface, generating powerful magnetizing fields impacting both native lunar rocks and extralunar projectile materials. We have reconstructed 3-D distribution maps of these ancient subsurface currents and developed coupling models of magnetic and electric fields that take into account the subsurface density in the prominent lunar maria and basins. Our simulations suggest these ancient currents could have reached density up to 13 A/m2, with surface magnetizing field as strong as 469 μT. We propose that these intense electrical current discharges in the crust originate from ancient interior dynamo activity.

astro-ph.EP↗

arXiv2024

Landmark-guided Diffusion Model for High-fidelity and Temporally Coherent Talking Head Generation

Audio-driven talking head generation is a significant and challenging task applicable to various fields such as virtual avatars, film production, and online conferences. However, the existing GAN-based models emphasize generating well-synchronized lip shapes but overlook the visual quality of generated frames, while diffusion-based models prioritize generating high-quality frames but neglect lip shape matching, resulting in jittery mouth movements. To address the aforementioned problems, we introduce a two-stage diffusion-based model. The first stage involves generating synchronized facial landmarks based on the given speech. In the second stage, these generated landmarks serve as a condition in the denoising process, aiming to optimize mouth jitter issues and generate high-fidelity, well-synchronized, and temporally coherent talking head videos. Extensive experiments demonstrate that our model yields the best performance.

cs.CV↗

arXiv2024

Integrated Sensing and Communication Under DISCO Physical-Layer Jamming Attacks

Integrated sensing and communication (ISAC) systems traditionally presuppose that sensing and communication (S&C) channels remain approximately constant during their coherence time. However, a "DISCO" reconfigurable intelligent surface (DRIS), i.e., an illegitimate RIS with random, time-varying reflection properties that acts like a "disco ball," introduces a paradigm shift that enables active channel aging more rapidly during the channel coherence time. In this letter, we investigate the impact of DISCO jamming attacks launched by a DRISbased fully-passive jammer (FPJ) on an ISAC system. Specifically, an ISAC problem formulation and a corresponding waveform optimization are presented in which the ISAC waveform design considers the trade-off between the S&C performance and is formulated as a Pareto optimization problem. Moreover, a theoretical analysis is conducted to quantify the impact of DISCO jamming attacks. Numerical results are presented to evaluate the S&C performance under DISCO jamming attacks and to validate the derived theoretical analysis.

eess.SP↗

arXiv2024

DISCO Might Not Be Funky: Random Intelligent Reflective Surface Configurations That Attack

Emerging intelligent reflective surfaces (IRSs) significantly improve system performance, but also pose a significant risk for physical layer security (PLS). Unlike the extensive research on legitimate IRS-enhanced communications, in this article we present an adversarial IRS-based fully-passive jammer (FPJ). We describe typical application scenarios for Disco IRS (DIRS)-based FPJ, where an illegitimate IRS with random, time-varying reflection properties acts like a "disco ball" to randomly change the propagation environment. We introduce the principles of DIRS-based FPJ and overview existing investigations of the technology, including a design example employing one-bit phase shifters. The DIRS-based FPJ can be implemented without either jamming power or channel state information (CSI) for the legitimate users (LUs). It does not suffer from the energy constraints of traditional active jammers, nor does it require any knowledge of the LU channels. In addition to the proposed jamming attack, we also propose an anti-jamming strategy that requires only statistical rather than instantaneous CSI. Furthermore, we present a data frame structure that enables the legitimate access point (AP) to estimate the DIRS-jammed channels' statistical characteristics in the presence of the DIRS jamming. Typical cases are discussed to show the impact of the DIRS-based FPJ and the feasibility of the anti-jamming precoder (AJP). Moreover, we outline future research directions and challenges for the DIRS-based FPJ and its anti-jamming precoding to stimulate this line of research and pave the way for practical applications.

eess.SP↗

arXiv2024

On Gradient-like Explanation under a Black-box Setting: When Black-box Explanations Become as Good as White-box

Attribution methods shed light on the explainability of data-driven approaches such as deep learning models by uncovering the most influential features in a to-be-explained decision. While determining feature attributions via gradients delivers promising results, the internal access required for acquiring gradients can be impractical under safety concerns, thus limiting the applicability of gradient-based approaches. In response to such limited flexibility, this paper presents \methodAbr~(gradient-estimation-based explanation), an approach that produces gradient-like explanations through only query-level access. The proposed approach holds a set of fundamental properties for attribution methods, which are mathematically rigorously proved, ensuring the quality of its explanations. In addition to the theoretical analysis, with a focus on image data, the experimental results empirically demonstrate the superiority of the proposed method over state-of-the-art black-box methods and its competitive performance compared to methods with full access.

cs.LG↗

arXiv2024

A Logical Pattern Memory Pre-trained Model for Entailment Tree Generation

Generating coherent and credible explanations remains a significant challenge in the field of AI. In recent years, researchers have delved into the utilization of entailment trees to depict explanations, which exhibit a reasoning process of how a hypothesis is deduced from the supporting facts. However, existing models often overlook the importance of generating intermediate conclusions with logical consistency from the given facts, leading to inaccurate conclusions and undermining the overall credibility of entailment trees. To address this limitation, we propose the logical pattern memory pre-trained model (LMPM). LMPM incorporates an external memory structure to learn and store the latent representations of logical patterns, which aids in generating logically consistent conclusions. Furthermore, to mitigate the influence of logically irrelevant domain knowledge in the Wikipedia-based data, we introduce an entity abstraction approach to construct the dataset for pre-training LMPM. The experimental results highlight the effectiveness of our approach in improving the quality of entailment tree generation. By leveraging logical entailment patterns, our model produces more coherent and reasonable conclusions that closely align with the underlying premises. Code and Data are released at https://github.com/YuanLi95/T5-LMPM

cs.CL↗

arXiv2024

Anti-Jamming Precoding Against Disco Intelligent Reflecting Surfaces Based Fully-Passive Jamming Attacks

Emerging intelligent reflecting surfaces (IRSs) significantly improve system performance, but also pose a huge risk for physical layer security. Existing works have illustrated that a disco IRS (DIRS), i.e., an illegitimate IRS with random time-varying reflection properties (like a "disco ball"), can be employed by an attacker to actively age the channels of legitimate users (LUs). Such active channel aging (ACA) generated by the DIRS can be employed to jam multi-user multiple-input single-output (MU-MISO) systems without relying on either jamming power or LU channel state information (CSI). To address the significant threats posed by DIRS-based fully-passive jammers (FPJs), an anti-jamming precoder is proposed that requires only the statistical characteristics of the DIRS-based ACA channels instead of their CSI. The statistical characteristics of DIRS-jammed channels are first derived, and then the anti-jamming precoder is derived based on the statistical characteristics. Furthermore, we prove that the anti-jamming precoder can achieve the maximum signal-to-jamming-plus-noise ratio (SJNR). To acquire the ACA statistics without changing the system architecture or cooperating with the illegitimate DIRS, we design a data frame structure that the legitimate access point (AP) can use to estimate the statistical characteristics. During the designed data frame, the LUs only need to feed back their received power to the legitimate AP when they detect jamming attacks. Numerical results are also presented to evaluate the effectiveness of the proposed anti-jamming precoder against the DIRS-based FPJs and the feasibility of the designed data frame used by the legitimate AP to estimate the statistical characteristics.

eess.SP↗

arXiv2024

IRS-Enhanced Anti-Jamming Precoding Against DISCO Physical Layer Jamming Attacks

Illegitimate intelligent reflective surfaces (IRSs) can pose significant physical layer security risks on multi-user multiple-input single-output (MU-MISO) systems. Recently, a DISCO approach has been proposed an illegitimate IRS with random and time-varying reflection coefficients, referred to as a "disco" IRS (DIRS). Such DIRS can attack MU-MISO systems without relying on either jamming power or channel state information (CSI), and classical anti-jamming techniques are ineffective for the DIRS-based fully-passive jammers (DIRS-based FPJs). In this paper, we propose an IRS-enhanced anti-jamming precoder against DIRS-based FPJs that requires only statistical rather than instantaneous CSI of the DIRS-jammed channels. Specifically, a legitimate IRS is introduced to reduce the strength of the DIRS-based jamming relative to the transmit signals at a legitimate user (LU). In addition, the active beamforming at the legitimate access point (AP) is designed to maximize the signal-to-jamming-plus-noise ratios (SJNRs). Numerical results are presented to evaluate the effectiveness of the proposed IRS-enhanced anti-jamming precoder against DIRS-based FPJs.

cs.IT↗

arXiv2023

Enhancing Large Language Models for Secure Code Generation: A Dataset-driven Study on Vulnerability Mitigation

Large language models (LLMs) have brought significant advancements to code generation, benefiting both novice and experienced developers. However, their training using unsanitized data from open-source repositories, like GitHub, introduces the risk of inadvertently propagating security vulnerabilities. To effectively mitigate this concern, this paper presents a comprehensive study focused on evaluating and enhancing code LLMs from a software security perspective. We introduce SecuCoGen\footnote{SecuCoGen has been uploaded as supplemental material and will be made publicly available after publication.}, a meticulously curated dataset targeting 21 critical vulnerability types. SecuCoGen comprises 180 samples and serves as the foundation for conducting experiments on three crucial code-related tasks: code generation, code repair and vulnerability classification, with a strong emphasis on security. Our experimental results reveal that existing models often overlook security concerns during code generation, leading to the generation of vulnerable code. To address this, we propose effective approaches to mitigate the security vulnerabilities and enhance the overall robustness of code generated by LLMs. Moreover, our study identifies weaknesses in existing models' ability to repair vulnerable code, even when provided with vulnerability information. Additionally, certain vulnerability types pose challenges for the models, hindering their performance in vulnerability classification. Based on these findings, we believe our study will have a positive impact on the software engineering community, inspiring the development of improved methods for training and utilizing LLMs, thereby leading to safer and more trustworthy model deployment.

cs.SE↗

arXiv2023

Advances in Atomic Time Scale imaging with a Fine Intrinsic Spatial Resolution

Atomic time scale imaging, opening a new era for studying dynamics in microcosmos, is presently attracting immense research interesting on the global level due to its powerful ability. On the atom level, physics, chemistry, and biology are identical for researching atom motion and atomic state change. The light possesses twoness, the information carrier and the research resource. The most fundamental principle of this imaging is that light records the event modulated light field by itself, so called all optical imaging. This paper can answer what is the essential standard to develop and evaluate atomic time scale imaging, what is the optimal imaging system, and what are the typical techniques to implement this imaging, up to now. At present, the best record in the experiment, made by multistage optical parametric amplification (MOPA), is realizing 50 fs resolved optical imaging with a spatial resolution of ~83 lp/mm at an effective framing rate of 10^13 fps for recording an ultrafast optical lattice with its rotating speed up to 10^13 rad/s.

physics.optics↗

arXiv2023

An Anti-Jamming Strategy for Disco Intelligent Reflecting Surfaces Based Fully-Passive Jamming Attacks

Emerging intelligent reflecting surfaces (IRSs) significantly improve system performance, while also pose a huge risk for physical layer security. A disco IRS (DIRS), i.e., an illegitimate IRS with random time-varying reflection properties, can be employed by an attacker to actively age the channels of legitimate users (LUs). Such active channel aging (ACA) generated by the DIRS-based fully-passive jammer (FPJ) can be applied to jam multi-user multiple-input single-output (MU-MISO) systems without relying on either jamming power or LU channel state information (CSI). To address the significant threats posed by the DIRS-based FPJ, an anti-jamming strategy is proposed that requires only the statistical characteristics of DIRS-jammed channels instead of their CSI. Statistical characteristics of DIRS-jammed channels are first derived, and then the anti-jamming precoder is given based on the derived statistical characteristics. Numerical results are also presented to evaluate the effectiveness of the proposed anti-jamming precoder against the DIRS-based FPJ.

eess.SP↗

arXiv2023

BiTimeBERT: Extending Pre-Trained Language Representations with Bi-Temporal Information

Time is an important aspect of documents and is used in a range of NLP and IR tasks. In this work, we investigate methods for incorporating temporal information during pre-training to further improve the performance on time-related tasks. Compared with common pre-trained language models like BERT which utilize synchronic document collections (e.g., BookCorpus and Wikipedia) as the training corpora, we use long-span temporal news article collection for building word representations. We introduce BiTimeBERT, a novel language representation model trained on a temporal collection of news articles via two new pre-training tasks, which harnesses two distinct temporal signals to construct time-aware language representations. The experimental results show that BiTimeBERT consistently outperforms BERT and other existing pre-trained models with substantial gains on different downstream NLP tasks and applications for which time is of importance (e.g., the accuracy improvement over BERT is 155\% on the event time estimation task).

cs.CL↗

arXiv2023

Disco Intelligent Reflecting Surfaces: Active Channel Aging for Fully-Passive Jamming Attacks

Due to the open communications environment in wireless channels, wireless networks are vulnerable to jamming attacks. However, existing approaches for jamming rely on knowledge of the legitimate users' (LUs') channels, extra jamming power, or both. To raise concerns about the potential threats posed by illegitimate intelligent reflecting surfaces (IRSs), we propose an alternative method to launch jamming attacks on LUs without either LU channel state information (CSI) or jamming power. The proposed approach employs an adversarial IRS with random phase shifts, referred to as a "disco" IRS (DIRS), that acts like a "disco ball" to actively age the LUs' channels. Such active channel aging (ACA) interference can be used to launch jamming attacks on multi-user multiple-input single-output (MU-MISO) systems. The proposed DIRS-based fully-passive jammer (FPJ) can jam LUs with no additional jamming power or knowledge of the LU CSI, and it can not be mitigated by classical anti-jamming approaches. A theoretical analysis of the proposed DIRS-based FPJ that provides an evaluation of the DIRS-based jamming attacks is derived. Based on this detailed theoretical analysis, some unique properties of the proposed DIRS-based FPJ can be obtained. Furthermore, a design example of the proposed DIRS-based FPJ based on one-bit quantization of the IRS phases is demonstrated to be sufficient for implementing the jamming attack. In addition, numerical results are provided to show the effectiveness of the derived theoretical analysis and the jamming impact of the proposed DIRS-based FPJ.

eess.SP↗

arXiv2023

Knowledge Graph Quality Evaluation under Incomplete Information

Knowledge graphs (KGs) have attracted more and more attentions because of their fundamental roles in many tasks. Quality evaluation for KGs is thus crucial and indispensable. Existing methods in this field evaluate KGs by either proposing new quality metrics from different dimensions or measuring performances at KG construction stages. However, there are two major issues with those methods. First, they highly rely on raw data in KGs, which makes KGs' internal information exposed during quality evaluation. Second, they consider more about the quality at data level instead of ability level, where the latter one is more important for downstream applications. To address these issues, we propose a knowledge graph quality evaluation framework under incomplete information (QEII). The quality evaluation task is transformed into an adversarial Q&A game between two KGs. Winner of the game is thus considered to have better qualities. During the evaluation process, no raw data is exposed, which ensures information protection. Experimental results on four pairs of KGs demonstrate that, compared with baselines, the QEII implements a reasonable quality evaluation at ability level under incomplete information.

cs.AI↗