arXiv subjects

Kai Tian

Publications and source records attributed to Kai Tian.

At least 37 records · Page 2Linked to original sources

arXiv2025

Code-Vision: Evaluating Multimodal LLMs Logic Understanding and Code Generation Capabilities

This paper introduces Code-Vision, a benchmark designed to evaluate the logical understanding and code generation capabilities of Multimodal Large Language Models (MLLMs). It challenges MLLMs to generate a correct program that fulfills specific functionality requirements based on a given flowchart, which visually represents the desired algorithm or process. Code-Vision comprises three subsets: HumanEval-V, Algorithm, and MATH, which evaluate MLLMs' coding abilities across basic programming, algorithmic, and mathematical problem-solving domains. Our experiments evaluate 12 MLLMs on Code-Vision. Experimental results demonstrate that there is a large performance difference between proprietary and open-source models. On Hard problems, GPT-4o can achieve 79.3% pass@1, but the best open-source model only achieves 15%. Further experiments reveal that Code-Vision can pose unique challenges compared to other multimodal reasoning benchmarks MMCode and MathVista. We also explore the reason for the poor performance of the open-source models. All data and codes are available at https://github.com/wanghanbinpanda/CodeVision.

cs.CL↗

arXiv2024

Exploring Adversarial Robustness of Deep State Space Models

Deep State Space Models (SSMs) have proven effective in numerous task scenarios but face significant security challenges due to Adversarial Perturbations (APs) in real-world deployments. Adversarial Training (AT) is a mainstream approach to enhancing Adversarial Robustness (AR) and has been validated on various traditional DNN architectures. However, its effectiveness in improving the AR of SSMs remains unclear. While many enhancements in SSM components, such as integrating Attention mechanisms and expanding to data-dependent SSM parameterizations, have brought significant gains in Standard Training (ST) settings, their potential benefits in AT remain unexplored. To investigate this, we evaluate existing structural variants of SSMs with AT to assess their AR performance. We observe that pure SSM structures struggle to benefit from AT, whereas incorporating Attention yields a markedly better trade-off between robustness and generalization for SSMs in AT compared to other components. Nonetheless, the integration of Attention also leads to Robust Overfitting (RO) issues. To understand these phenomena, we empirically and theoretically analyze the output error of SSMs under AP. We find that fixed-parameterized SSMs have output error bounds strictly related to their parameters, limiting their AT benefits, while input-dependent SSMs may face the problem of error explosion. Furthermore, we show that the Attention component effectively scales the output error of SSMs during training, enabling them to benefit more from AT, but at the cost of introducing RO due to its high model complexity. Inspired by this, we propose a simple and effective Adaptive Scaling (AdS) mechanism that brings AT performance close to Attention-integrated SSMs without introducing the issue of RO. Our code is available at https://github.com/Biqing-Qi/Exploring-Adversarial-Robustness-of-Deep-State-Space-Models.git.

cs.LG↗

arXiv2024

Large Language Models as Biomedical Hypothesis Generators: A Comprehensive Evaluation

The rapid growth of biomedical knowledge has outpaced our ability to efficiently extract insights and generate novel hypotheses. Large language models (LLMs) have emerged as a promising tool to revolutionize knowledge interaction and potentially accelerate biomedical discovery. In this paper, we present a comprehensive evaluation of LLMs as biomedical hypothesis generators. We construct a dataset of background-hypothesis pairs from biomedical literature, carefully partitioned into training, seen, and unseen test sets based on publication date to mitigate data contamination. Using this dataset, we assess the hypothesis generation capabilities of top-tier instructed models in zero-shot, few-shot, and fine-tuning settings. To enhance the exploration of uncertainty, a crucial aspect of scientific discovery, we incorporate tool use and multi-agent interactions in our evaluation framework. Furthermore, we propose four novel metrics grounded in extensive literature review to evaluate the quality of generated hypotheses, considering both LLM-based and human assessments. Our experiments yield two key findings: 1) LLMs can generate novel and validated hypotheses, even when tested on literature unseen during training, and 2) Increasing uncertainty through multi-agent interactions and tool use can facilitate diverse candidate generation and improve zero-shot hypothesis generation performance. However, we also observe that the integration of additional knowledge through few-shot learning and tool use may not always lead to performance gains, highlighting the need for careful consideration of the type and scope of external knowledge incorporated. These findings underscore the potential of LLMs as powerful aids in biomedical hypothesis generation and provide valuable insights to guide further research in this area.

cs.CL↗

arXiv2024

Not all explicit cues help communicate: Pedestrians' perceptions, fixations, and decisions toward automated vehicles with varied appearance

Given pedestrians' vulnerability in road traffic, it remains unclear how novel AV appearances will impact pedestrians crossing behaviour. To address this gap, this study pioneers an investigation into the influence of AVs' exterior design, correlated with their kinematics, on pedestrians' road-crossing perception and decision-making. A video-based eye-tracking experimental study was conducted with 61 participants who responded to video stimuli depicting a manipulated vehicle approaching a predefined road-crossing location on an unsignalized, two-way road. The vehicle's kinematic pattern was manipulated into yielding and non-yielding, and its external appearances were varied across five types: with a human driver (as a conventional vehicle), with no driver (as an AV), with text-based identity indications, with roof radar sensors, with dynamic eHMIs adjusted to vehicle kinematics. Participants' perceived clarity, crossing initiation distance (CID), crossing decision time (CDT), and gaze behaviour, during interactions were recorded and reported. The results indicated that AVs' kinematic profiles play a dominant role in pedestrians' road-crossing decisions, supported by their subjective evaluations, CID, CDT, and gaze patterns during interactions. Moreover, the use of clear eHMI, such as dynamic pedestrian icons, reduced pedestrians' visual load, enhanced their perceptual clarity, expedited road-crossing decisions, and thereby improved overall crossing efficiency. However, it was found that both textual identity indications and roof radar sensors have no significant effect on pedestrians' decisions but negatively impact pedestrians' visual attention, as evidenced by heightened fixation counts and prolonged fixation durations, particularly under yielding conditions. Excessive visual and cognitive resource occupation suggests that not all explicit cues facilitate human-vehicle communication.

cs.HC↗

arXiv2024

Nonlocal Symmetries of Two 2-component Equations of Camassa-Holm Type

For a 2-component Camassa-Holm equation, as well as a 2-component generalization of the modified Camassa-Holm equation, nonlocal infinitesimal symmetries quadratically depending on eigenfunctions of linear spectral problems are constructed from functional gradients of spectral parameters. With appropriate pseudo-potentials, these nonlocal infinitesimal symmetries are prolonged to enlarged systems, and then explicitly integrated to generate symmetry transformations in finite form for enlarged systems. As implementations of these finite symmetry transformations, some kinds of nontrivial solutions and Bäcklund transformations are derived for both equations.

nlin.SI↗

arXiv2023

Large Language Models are Zero Shot Hypothesis Proposers

Significant scientific discoveries have driven the progress of human civilisation. The explosion of scientific literature and data has created information barriers across disciplines that have slowed the pace of scientific discovery. Large Language Models (LLMs) hold a wealth of global and interdisciplinary knowledge that promises to break down these information barriers and foster a new wave of scientific discovery. However, the potential of LLMs for scientific discovery has not been formally explored. In this paper, we start from investigating whether LLMs can propose scientific hypotheses. To this end, we construct a dataset consist of background knowledge and hypothesis pairs from biomedical literature. The dataset is divided into training, seen, and unseen test sets based on the publication date to control visibility. We subsequently evaluate the hypothesis generation capabilities of various top-tier instructed models in zero-shot, few-shot, and fine-tuning settings, including both closed and open-source LLMs. Additionally, we introduce an LLM-based multi-agent cooperative framework with different role designs and external tools to enhance the capabilities related to generating hypotheses. We also design four metrics through a comprehensive review to evaluate the generated hypotheses for both ChatGPT-based and human evaluations. Through experiments and analyses, we arrive at the following findings: 1) LLMs surprisingly generate untrained yet validated hypotheses from testing literature. 2) Increasing uncertainty facilitates candidate generation, potentially enhancing zero-shot hypothesis generation capabilities. These findings strongly support the potential of LLMs as catalysts for new scientific discoveries and guide further exploration.

cs.CL↗

arXiv2023

Social Interaction-Aware Dynamical Models and Decision Making for Autonomous Vehicles

Interaction-aware Autonomous Driving (IAAD) is a rapidly growing field of research that focuses on the development of autonomous vehicles (AVs) that are capable of interacting safely and efficiently with human road users. This is a challenging task, as it requires the autonomous vehicle to be able to understand and predict the behaviour of human road users. In this literature review, the current state of IAAD research is surveyed in this work. Commencing with an examination of terminology, attention is drawn to challenges and existing models employed for modelling the behaviour of drivers and pedestrians. Next, a comprehensive review is conducted on various techniques proposed for interaction modelling, encompassing cognitive methods, machine learning approaches, and game-theoretic methods. The conclusion is reached through a discussion of potential advantages and risks associated with IAAD, along with the illumination of pivotal research inquiries necessitating future exploration.

cs.HC↗

arXiv2023

Deconstructing Pedestrian Crossing Decision-making in Interactions with Continuous Traffic: an Anthropomorphic Model

As safe and comfortable interactions with pedestrians could contribute to automated vehicles' (AVs) social acceptance and scale, increasing attention has been drawn to computational pedestrian behavior models. However, very limited studies characterize pedestrian crossing behavior based on specific behavioral mechanisms, as those mechanisms underpinning pedestrian road behavior are not yet clear. Here, we reinterpret pedestrian crossing behavior based on a deconstructed crossing decision process at uncontrolled intersections with continuous traffic. Notably, we explain and model pedestrian crossing behavior as they wait for crossing opportunities, optimizing crossing decisions by comparing the visual collision risk of approaching vehicles around them. A collision risk-based crossing initiation model is proposed to characterize the time-dynamic nature of pedestrian crossing decisions. A simulation tool is established to reproduce pedestrian behavior by employing the proposed model and a social force model. Two datasets collected in a CAVE-based immersive pedestrian simulator are applied to calibrate and validate the model. The model predicts pedestrian crossing decisions across all traffic scenarios well. In particular, by considering the decision strategy that pedestrians compare the collision risk of surrounding traffic gaps, model performance is significantly improved. Moreover, the collision risk-based crossing initiation model accurately captures the timing of pedestrian crossing initiations within each gap. This work concisely demonstrates how pedestrians dynamically adapt their crossings in continuous traffic based on perceived collision risk, potentially providing insights into modeling coupled human-AV interactions or serving as a tool to realize human-like pedestrian road behavior in virtual AVs test platforms.

stat.AP↗

arXiv2022

TED: Towards Discovering Top-k Edge-Diversified Patterns in a Graph Database

With an exponentially growing number of graphs from disparate repositories, there is a strong need to analyze a graph database containing an extensive collection of small- or medium-sized data graphs (e.g., chemical compounds). Although subgraph enumeration and subgraph mining have been proposed to bring insights into a graph database by a set of subgraph structures, they often end up with similar or homogenous topologies, which is undesirable in many graph applications. To address this limitation, we propose the Top-k Edge-Diversified Patterns Discovery problem to retrieve a set of subgraphs that cover the maximum number of edges in a database. To efficiently process such query, we present a generic and extensible framework called Ted which achieves a guaranteed approximation ratio to the optimal result. Two optimization strategies are further developed to improve the performance. Experimental studies on real-world datasets demonstrate the superiority of Ted to traditional techniques.

cs.DB↗

arXiv2019

Non-Local ConvLSTM for Video Compression Artifact Reduction

Video compression artifact reduction aims to recover high-quality videos from low-quality compressed videos. Most existing approaches use a single neighboring frame or a pair of neighboring frames (preceding and/or following the target frame) for this task. Furthermore, as frames of high quality overall may contain low-quality patches, and high-quality patches may exist in frames of low quality overall, current methods focusing on nearby peak-quality frames (PQFs) may miss high-quality details in low-quality frames. To remedy these shortcomings, in this paper we propose a novel end-to-end deep neural network called non-local ConvLSTM (NL-ConvLSTM in short) that exploits multiple consecutive frames. An approximate non-local strategy is introduced in NL-ConvLSTM to capture global motion patterns and trace the spatiotemporal dependency in a video sequence. This approximate strategy makes the non-local module work in a fast and low space-cost way. Our method uses the preceding and following frames of the target frame to generate a residual, from which a higher quality frame is reconstructed. Experiments on two datasets show that NL-ConvLSTM outperforms the existing methods.

eess.IV↗

arXiv2019

Learning Competitive and Discriminative Reconstructions for Anomaly Detection

Most of the existing methods for anomaly detection use only positive data to learn the data distribution, thus they usually need a pre-defined threshold at the detection stage to determine whether a test instance is an outlier. Unfortunately, a good threshold is vital for the performance and it is really hard to find an optimal one. In this paper, we take the discriminative information implied in unlabeled data into consideration and propose a new method for anomaly detection that can learn the labels of unlabelled data directly. Our proposed method has an end-to-end architecture with one encoder and two decoders that are trained to model inliers and outliers' data distributions in a competitive way. This architecture works in a discriminative manner without suffering from overfitting, and the training algorithm of our model is adopted from SGD, thus it is efficient and scalable even for large-scale datasets. Empirical studies on 7 datasets including KDD99, MNIST, Caltech-256, and ImageNet etc. show that our model outperforms the state-of-the-art methods.

cs.LG↗

arXiv2018

Global Semantic Consistency for Zero-Shot Learning

In image recognition, there are many cases where training samples cannot cover all target classes. Zero-shot learning (ZSL) utilizes the class semantic information to classify samples of the unseen categories that have no corresponding samples contained in the training set. In this paper, we propose an end-to-end framework, called Global Semantic Consistency Network (GSC-Net for short), which makes complete use of the semantic information of both seen and unseen classes, to support effective zero-shot learning. We also adopt a soft label embedding loss to further exploit the semantic relationships among classes. To adapt GSC-Net to a more practical setting, Generalized Zero-shot Learning (GZSL), we introduce a parametric novelty detection mechanism. Our approach achieves the state-of-the-art performance on both ZSL and GZSL tasks over three visual attribute datasets, which validates the effectiveness and advantage of the proposed framework.

cs.CV↗

arXiv2017

Ground-to-satellite quantum teleportation

An arbitrary unknown quantum state cannot be precisely measured or perfectly replicated. However, quantum teleportation allows faithful transfer of unknown quantum states from one object to another over long distance, without physical travelling of the object itself. Long-distance teleportation has been recognized as a fundamental element in protocols such as large-scale quantum networks and distributed quantum computation. However, the previous teleportation experiments between distant locations were limited to a distance on the order of 100 kilometers, due to photon loss in optical fibres or terrestrial free-space channels. An outstanding open challenge for a global-scale "quantum internet" is to significantly extend the range for teleportation. A promising solution to this problem is exploiting satellite platform and space-based link, which can conveniently connect two remote points on the Earth with greatly reduced channel loss because most of the photons' propagation path is in empty space. Here, we report the first quantum teleportation of independent single-photon qubits from a ground observatory to a low Earth orbit satellite - through an up-link channel - with a distance up to 1400 km. To optimize the link efficiency and overcome the atmospheric turbulence in the up-link, a series of techniques are developed, including a compact ultra-bright source of multi-photon entanglement, narrow beam divergence, high-bandwidth and high-accuracy acquiring, pointing, and tracking (APT). We demonstrate successful quantum teleportation for six input states in mutually unbiased bases with an average fidelity of 0.80+/-0.01, well above the classical limit. This work establishes the first ground-to-satellite up-link for faithful and ultra-long-distance quantum teleportation, an essential step toward global-scale quantum internet.

quant-ph↗

arXiv2016

Symbolic Representation and Classification of N=1 Supersymmetric Evolutionary Equations

We extend the symbolic representation to the ring of N=1 supersymmetric differential polynomials, and demonstrate that operations on the ring, such as the super derivative, Frechet derivative and super commutator, can be carried out in the symbolic way. Using the symbolic representation, we classify scalar $λ$-homogeneous N=1 supersymmetric evolutionary equations with nonzero linear term when $λ>0$ for arbitrary order and give a comprehensive description of all such integrable equations.

nlin.SI↗

arXiv2015

Conservation laws and symmetries of Hunter-Saxton equation: revisited

Through a reciprocal transformation $\mathcal{T}_0$ induced by the conservation law $\partial_t(u_x^2) = \partial_x(2uu_x^2)$, the Hunter-Saxton (HS) equation $u_{xt} = 2uu_{2x} + u_x^2$ is shown to possess conserved densities involving arbitrary smooth functions, which have their roots in infinitesimal symmetries of $w_t = w^2$, the counterpart of the HS equation under $\mathcal{T}_0$. Hierarchies of commuting symmetries of the HS equation are studied under appropriate changes of variables initiated by $\mathcal{T}_0$, and two of these are linearized while the other is identical to the hierarchy of commuting symmetries admitted by the potential modified Korteweg-de Vries equation. A fifth order symmetry of the HS equation is endowed with a sixth order hereditary recursion operator by its connection with the Fordy-Gibbons equation. These results reveal the origin for the rich and remarkable structures of the HS equation and partially answer the questions raised by Wang [{\it Nonlinearity} {\bf 23}(2010) 2009].

nlin.SI↗

arXiv2011

A non-standard Lax formulation of the Harry Dym hierarchy and its supersymmetric extension

For the Harry Dym hierarchy, a non-standard Lax formulation is deduced from that of Korteweg-de Vries (KdV) equation through a reciprocal transformation. By supersymmetrizing this Lax operator, a new N=2 supersymmetric extension of the Harry Dym hierarchy is constructed, and is further shown to be linked to one of the N=2 supersymmetric KdV equations through superconformal transformation. The bosonic limit of this new N=2 supersymmetric Harry Dym equation is related to a coupled system of KdV-MKdV equations.

nlin.SI↗

arXiv2011

The transformations between N=2 supersymmetric KdV and HD equations

The N=2 supercomformal transformations are employed to study supersymmetric integrable systems. It is proved that two known N=2 supersymmetric Harry Dym equations are transformed into two N=2 supersymmetric modified Kortweg-de Vries equations, thus are connected with two N=2 supersymmetric Kortweg-de Vries equations.

nlin.SI↗

arXiv2010

The Even and Odd Supersymmetric Hunter - Saxton and Liouville Equations

It is shown that two different supersymmetric extensions of the Harry Dym equation lead to two different negative hierarchies of the supersymmetric integrable equations. While the first one yields the known even supersymmetric Hunter - Saxton equation, the second one is a new odd supersymmetric Hunter - Saxton equation. It is further proved that these two supersymmetric extensions of the Hunter - Saxton equation are reciprocally transformed to two different supersymmetric extensions of the Liouville equation.

nlin.SI↗