arXiv subjects

Jiaxin Liu

Publications and source records attributed to Jiaxin Liu.

At least 73 records · Page 4Linked to original sources

arXiv2025

SQLFixAgent: Towards Semantic-Accurate Text-to-SQL Parsing via Consistency-Enhanced Multi-Agent Collaboration

While fine-tuned large language models (LLMs) excel in generating grammatically valid SQL in Text-to-SQL parsing, they often struggle to ensure semantic accuracy in queries, leading to user confusion and diminished system usability. To tackle this challenge, we introduce SQLFixAgent, a new consistency-enhanced multi-agent collaborative framework designed for detecting and repairing erroneous SQL. Our framework comprises a core agent, SQLRefiner, alongside two auxiliary agents: SQLReviewer and QueryCrafter. The SQLReviewer agent employs the rubber duck debugging method to identify potential semantic mismatches between SQL and user query. If the error is detected, the QueryCrafter agent generates multiple SQL as candidate repairs using a fine-tuned SQLTool. Subsequently, leveraging similar repair retrieval and failure memory reflection, the SQLRefiner agent selects the most fitting SQL statement from the candidates as the final repair. We evaluated our proposed framework on five Text-to-SQL benchmarks. The experimental results show that our method consistently enhances the performance of the baseline model, specifically achieving an execution accuracy improvement of over 3% on the Bird benchmark. Our framework also has a higher token efficiency compared to other advanced methods, making it more competitive.

cs.CL↗

arXiv2025

FairACE: Achieving Degree Fairness in Graph Neural Networks via Contrastive and Adversarial Group-Balanced Training

Fairness has been a significant challenge in graph neural networks (GNNs) since degree biases often result in un-equal prediction performance among nodes with varying degrees. Existing GNN models focus on prediction accuracy, frequently overlooking fairness across different degree groups. To addressthis issue, we propose a novel GNN framework, namely Fairness- Aware Asymmetric Contrastive Ensemble (FairACE), which inte-grates asymmetric contrastive learning with adversarial training to improve degree fairness. FairACE captures one-hop local neighborhood information and two-hop monophily similarity to create fairer node representations and employs a degree fairness regulator to balance performance between high-degree and low-degree nodes. During model training, a novel group-balanced fairness loss is proposed to minimize classification disparities across degree groups. In addition, we also propose a novel fairness metric, the Accuracy Distribution Gap (ADG), which can quantitatively assess and ensure equitable performance across different degree-based node groups. Experimental results on both synthetic and real-world datasets demonstrate that FairACE significantly improves degree fairness metrics while maintaining competitive accuracy in comparison to the state-of-the-art GNN models.

cs.LG↗

arXiv2025

Probing intensity noise in ultrafast pulses using the dispersive Fourier transform augmented by quantum sensitivity analysis

To reach the next frontier in multimode nonlinear optics, it is crucial to better understand the classical and quantum phenomena of systems with many interacting degrees of freedom -- both how they emerge and how they can be tailored to emerging applications, from multimode quantum light generation to optical computing. Soliton fission and Raman scattering comprise two such phenomena that are ideal testbeds for exploring multimode nonlinear optics, especially power-dependent physics. To fully capture the complexity of such processes, an experimental measurement technique capable of measuring shot-to-shot pulse variations is necessary. The dispersive Fourier transform (DFT) is the ideal technique to achieve this goal, using chromatic dispersion to temporally stretch an ultrafast pulse and map its spectrum onto a measurable temporal waveform. Here, we apply DFT to explore the power-dependent mean field and noise properties of soliton fission and Raman scattering. To explain quantum noise properties, the traditional approach is to perform several hundred stochastic simulations for computing statistics. In our work, we apply quantum sensitivity analysis (QSA) to compute the noise in any output observable based on fluctuations in the input pulse, all using a single backwards differentiation step. We find that the combination of DFT and QSA provides a powerful framework for understanding the quantum and classical properties of soliton fission and Raman scattering, and can be generalized to other multimode nonlinear phenomena.

physics.optics↗

arXiv2025

Multi-models with averaging in feature domain for non-invasive blood glucose estimation

Diabetes is a serious chronic metabolic disease. In the recent years, more and more consumer technology enterprises focusing on human health are committed to implementing accurate and non-invasive blood glucose algorithm in their products. However, due to the interference from the external environment, these wearable non-invasive methods yield the low estimation accuracy. To address this issue, this paper employs different models based on different ranges of the blood glucose values for performing the blood glucose estimation. First the photoplethysmograms (PPGs) are acquired and they are denoised via the bit plane singular spectrum analysis (SSA) method. Second, the features are extracted. For the data in the training set, first the features are averaged across the measurements in the feature domain via the optimization approach. Second, the random forest is employed to sort the importance of each feature. Third, the training set is divided into three subsets according to the reference blood glucose values. Fourth, the feature vectors and the corresponding blood glucose values in the same group are employed to build an individual model. Fifth, for each feature, the average of the feature values for all the measurements in the same subset is computed. For the data in the test set, first, the sum of the weighted distances between the test feature values and the average values obtained in the above is computed for each model. Here, the weights are defined based on the importance sorted by the random forest obtained in the above. The model corresponding to the smallest sum is assigned. Finally, the blood glucose value is estimated based on the corresponding model. Compared to the state of arts methods, our proposed method can effectively improve the estimation accuracy.

cs.LG↗

arXiv2025

TrajLLM: A Modular LLM-Enhanced Agent-Based Framework for Realistic Human Trajectory Simulation

This work leverages Large Language Models (LLMs) to simulate human mobility, addressing challenges like high costs and privacy concerns in traditional models. Our hierarchical framework integrates persona generation, activity selection, and destination prediction, using real-world demographic and psychological data to create realistic movement patterns. Both physical models and language models are employed to explore and demonstrate different methodologies for human mobility simulation. By structuring data with summarization and weighted density metrics, the system ensures scalable memory management while retaining actionable insights. Preliminary results indicate that LLM-driven simulations align with observed real-world patterns, offering scalable, interpretable insights for social problems such as urban planning, traffic management, and public health. The framework's ability to dynamically generate personas and activities enables it to provide adaptable and realistic daily routines. This study demonstrates the transformative potential of LLMs in advancing mobility modeling for societal and urban applications. The source code and interactive demo for our framework are available at https://github.com/cju0/TrajLLM.

cs.AI↗

arXiv2024

3D Multi-Object Tracking with Semi-Supervised GRU-Kalman Filter

3D Multi-Object Tracking (MOT), a fundamental component of environmental perception, is essential for intelligent systems like autonomous driving and robotic sensing. Although Tracking-by-Detection frameworks have demonstrated excellent performance in recent years, their application in real-world scenarios faces significant challenges. Object movement in complex environments is often highly nonlinear, while existing methods typically rely on linear approximations of motion. Furthermore, system noise is frequently modeled as a Gaussian distribution, which fails to capture the true complexity of the noise dynamics. These oversimplified modeling assumptions can lead to significant reductions in tracking precision. To address this, we propose a GRU-based MOT method, which introduces a learnable Kalman filter into the motion module. This approach is able to learn object motion characteristics through data-driven learning, thereby avoiding the need for manual model design and model error. At the same time, to avoid abnormal supervision caused by the wrong association between annotations and trajectories, we design a semi-supervised learning strategy to accelerate the convergence speed and improve the robustness of the model. Evaluation experiment on the nuScenes and Argoverse2 datasets demonstrates that our system exhibits superior performance and significant potential compared to traditional TBD methods.

cs.RO↗

arXiv2024

LLM-Measure: Generating Valid, Consistent, and Reproducible Text-Based Measures for Social Science Research

The increasing use of text as data in social science research necessitates the development of valid, consistent, reproducible, and efficient methods for generating text-based concept measures. This paper presents a novel method that leverages the internal hidden states of large language models (LLMs) to generate these concept measures. Specifically, the proposed method learns a concept vector that captures how the LLM internally represents the target concept, then estimates the concept value for text data by projecting the text's LLM hidden states onto the concept vector. Three replication studies demonstrate the method's effectiveness in producing highly valid, consistent, and reproducible text-based measures across various social science research contexts, highlighting its potential as a valuable tool for the research community.

cs.CL↗

arXiv2024

The Modified Combo i3+3 Design for Novel-Novel Combination Dose-Finding Trials in Oncology

We consider a modified Ci3+3 (MCi3+3) design for dual-agent dose-finding trials in which both agents are tested on multiple doses. This usually happens when the agents are novel therapies. The MCi3+3 design offers a two-stage or three-stage version, depending on the practical need. The first stage begins with single-agent dose escalation, the second stage launches a model-free combination dose finding for both agents, and optionally, the third stage follows with a model-based design. MCi3+3 aims to maintain a relatively simple framework to facilitate practical application, while also address challenges that are unique to novel-novel combination dose finding. Through simulations, we demonstrate that the MCi3+3 design adeptly manages various toxicity scenarios. It exhibits operational characteristics on par with other combination designs, while offering an enhanced safety profile. The design is motivated and tested for a real-life clinical trial.

stat.AP↗

arXiv2024

MS23D: A 3D Object Detection Method Using Multi-Scale Semantic Feature Points to Construct 3D Feature Layer

LiDAR point clouds can effectively depict the motion and posture of objects in three-dimensional space. Many studies accomplish the 3D object detection by voxelizing point clouds. However, in autonomous driving scenarios, the sparsity and hollowness of point clouds create some difficulties for voxel-based methods. The sparsity of point clouds makes it challenging to describe the geometric features of objects. The hollowness of point clouds poses difficulties for the aggregation of 3D features. We propose a two-stage 3D object detection framework, called MS23D. (1) We propose a method using voxel feature points from multi-branch to construct the 3D feature layer. Using voxel feature points from different branches, we construct a relatively compact 3D feature layer with rich semantic features. Additionally, we propose a distance-weighted sampling method, reducing the loss of foreground points caused by downsampling and allowing the 3D feature layer to retain more foreground points. (2) In response to the hollowness of point clouds, we predict the offsets between deep-level feature points and the object's centroid, making them as close as possible to the object's centroid. This enables the aggregation of these feature points with abundant semantic features. For feature points from shallow-level, we retain them on the object's surface to describe the geometric features of the object. To validate our approach, we evaluated its effectiveness on both the KITTI and ONCE datasets.

cs.CV↗

arXiv2024

A 65nm 8b-Activation 8b-Weight SRAM-Based Charge-Domain Computing-in-Memory Macro Using A Fully-Parallel Analog Adder Network and A Single-ADC Interface

Performing data-intensive tasks in the von Neumann architecture is challenging to achieve both high performance and power efficiency due to the memory wall bottleneck. Computing-in-memory (CiM) is a promising mitigation approach by enabling parallel in-situ multiply-accumulate (MAC) operations within the memory with support from the peripheral interface and datapath. SRAM-based charge-domain CiM (CD-CiM) has shown its potential of enhanced power efficiency and computing accuracy. However, existing SRAM-based CD-CiM faces scaling challenges to meet the throughput requirement of high-performance multi-bit-quantization applications. This paper presents an SRAM-based high-throughput ReLU-optimized CD-CiM macro. It is capable of completing MAC and ReLU of two signed 8b vectors in one CiM cycle with only one A/D conversion. Along with non-linearity compensation for the analog computing and A/D conversion interfaces, this work achieves 51.2GOPS throughput and 10.3TOPS/W energy efficiency, while showing 88.6% accuracy in the CIFAR-10 dataset.

cs.AR↗

arXiv2024

Beyond Surface Similarity: Detecting Subtle Semantic Shifts in Financial Narratives

In this paper, we introduce the Financial-STS task, a financial domain-specific NLP task designed to measure the nuanced semantic similarity between pairs of financial narratives. These narratives originate from the financial statements of the same company but correspond to different periods, such as year-over-year comparisons. Measuring the subtle semantic differences between these paired narratives enables market stakeholders to gauge changes over time in the company's financial and operational situations, which is critical for financial decision-making. We find that existing pretrained embedding models and LLM embeddings fall short in discerning these subtle financial narrative shifts. To address this gap, we propose an LLM-augmented pipeline specifically designed for the Financial-STS task. Evaluation on a human-annotated dataset demonstrates that our proposed method outperforms existing methods trained on classic STS tasks and generic LLM embeddings.

cs.CL↗

arXiv2024

A Spatial Calibration Method for Robust Cooperative Perception

Cooperative perception is a promising technique for intelligent and connected vehicles through vehicle-to-everything (V2X) cooperation, provided that accurate pose information and relative pose transforms are available. Nevertheless, obtaining precise positioning information often entails high costs associated with navigation systems. {Hence, it is required to calibrate relative pose information for multi-agent cooperative perception.} This paper proposes a simple but effective object association approach named context-based matching (CBM), which identifies inter-agent object correspondences using intra-agent geometrical context. In detail, this method constructs contexts using the relative position of the detected bounding boxes, followed by local context matching and global consensus maximization. The optimal relative pose transform is estimated based on the matched correspondences, followed by cooperative perception fusion. Extensive experiments are conducted on both the simulated and real-world datasets. Even with larger inter-agent localization errors, high object association precision and decimeter-level relative pose calibration accuracy are achieved among the cooperating agents.

cs.RO↗

arXiv2024

Pharmacometrics-Enabled DOse OPtimization (PEDOOP) for Seamless Phase I-II Trials in Oncology

We consider a dose-optimization design for first-in-human oncology trial that aims to identify a suitable dose for late-phase drug development. The proposed approach, called the Pharmacometrics-Enabled DOse OPtimization (PEDOOP) design, incorporates observed patient-level pharmacokinetics (PK) measurements and latent pharmacodynamics (PD) information for trial decision making and dose optimization. PEDOOP consists of two seamless phases. In phase I, patient-level time-course drug concentrations, derived PD effects, and the toxicity outcomes from patients are integrated into a statistical model to estimate the dose-toxicity response. A simple dose-finding design guides dose escalation in phase I. At the end of the phase I dose finding, a graduation rule is used to assess the safety and efficacy of all the doses and select those with promising efficacy and acceptable safety for a randomized comparison against a control arm in phase II. In phase II, patients are randomized to the selected doses based on a fixed or adaptive randomization ratio. At the end of phase II, an optimal biological dose (OBD) is selected for late-phase development. We conduct simulation studies to assess the PEDOOP design in comparison to an existing seamless design that also combines phases I and II in a single trial.

stat.AP↗

arXiv2024

The Backfill i3+3 Design for Dose-Finding Trials in Oncology

We consider a formal statistical design that allows simultaneous enrollment of a main cohort and a backfill cohort of patients in a dose-finding trial. The goal is to accumulate more information at various doses to facilitate dose optimization. The proposed design, called Bi3+3, combines the simple dose-escalation algorithm in the i3+3 design and a model-based inference under the framework of probability of decisions (POD), both previously published. As a result, Bi3+3 provides a simple algorithm for backfilling patients to lower doses in a dose-finding trial once these doses exhibit safety profile in patients. The POD framework allows dosing decisions to be made when some backfill patients are still being followed with incomplete toxicity outcomes, thereby potentially expediting the clinical trial. At the end of the trial, Bi3+3 uses both toxicity and efficacy outcomes to estimate an optimal biological dose (OBD). The proposed inference is based on a dose-response model that takes into account either a monotone or plateau dose-efficacy relationship, which are frequently encountered in modern oncology drug development. Simulation studies show promising operating characteristics of the Bi3+3 design in comparison to existing designs.

stat.AP↗

arXiv2023

On the generation of attosecond gigawatt soft X-ray pulses through coherent Thomson backscattering

Collision between relativistic electron sheets and counter-propagating laser pulses is recognized as a promising way to produce intense attosecond X-rays through coherent Thomson backscattering (TBS). In a double-layer scheme, the electrons in an ultrathin solid foil are first pushed out by an intense laser driver and then interact with the laser reflected off a second foil to form a high-density relativistic electron sheet with vanishing transverse momentum. However, the repulsion between these concentrated electrons can increase the thickness of the layer, reducing both its density and subsequently the coherent TBS. Here, we present a systematic study on the evolution of the flying electron layer and find that its resulting thickness is determined by the interplay between the intrinsic space-charge expansion and the velocity compression induced by the drive laser. How the laser driver, the target areal density, the reflector and the collision laser intensity affect the properties of the produced X-rays is explored. Multi-dimensional particle-in-cell simulations indicate that employing this scheme in the nonlinear regime has the potential to stably produce soft X-rays with several GW peak power in hundreds of TW ultrafast laser facilities. The pulse duration can be tuned to tens of attoseconds. This compact and intense attosecond X-ray source may have broad applications in attosecond science.

physics.plasm-ph↗

arXiv2023

Fast and reconfigurable sort-in-memory system enabled by memristors

Sorting is fundamental and ubiquitous in modern computing systems. Hardware sorting systems are built based on comparison operations with Von Neumann architecture, but their performance are limited by the bandwidth between memory and comparison units and the performance of complementary metal-oxide-semiconductor (CMOS) based circuitry. Sort-in-memory (SIM) based on emerging memristors is desired but not yet available due to comparison operations that are challenging to be implemented within memristive memory. Here we report fast and reconfigurable SIM system enabled by digit read (DR) on 1-transistor-1-resistor (1T1R) memristor arrays. We develop DR tree node skipping (TNS) that support variable data quantity and data types, and extend TNS with multi-bank, bit-slice and multi-level strategies to enable cross-array TNS (CA-TNS) for practical adoptions. Experimented on benchmark sorting datasets, our memristor-enabled SIM system presents up to 3.32x~7.70x speedup, 6.23x~183.5x energy efficiency improvement and 2.23x~7.43x area reduction compared with state-of-the-art sorting systems. We apply such SIM system for shortest path search with Dijkstra's algorithm and neural network inference with in-situ pruning, demonstrating the capability in solving practical sorting tasks and the compatibility in integrating with other compute-in-memory (CIM) schemes. The comparison-free TNS/CA-TNS SIM enabled by memristors pushes sorting into a new paradigm of sort-in-memory for next-generation sorting systems.

cs.AR↗

arXiv2023

RetouchingFFHQ: A Large-scale Dataset for Fine-grained Face Retouching Detection

The widespread use of face retouching filters on short-video platforms has raised concerns about the authenticity of digital appearances and the impact of deceptive advertising. To address these issues, there is a pressing need to develop advanced face retouching techniques. However, the lack of large-scale and fine-grained face retouching datasets has been a major obstacle to progress in this field. In this paper, we introduce RetouchingFFHQ, a large-scale and fine-grained face retouching dataset that contains over half a million conditionally-retouched images. RetouchingFFHQ stands out from previous datasets due to its large scale, high quality, fine-grainedness, and customization. By including four typical types of face retouching operations and different retouching levels, we extend the binary face retouching detection into a fine-grained, multi-retouching type, and multi-retouching level estimation problem. Additionally, we propose a Multi-granularity Attention Module (MAM) as a plugin for CNN backbones for enhanced cross-scale representation learning. Extensive experiments using different baselines as well as our proposed method on RetouchingFFHQ show decent performance on face retouching detection. With the proposed new dataset, we believe there is great potential for future work to tackle the challenging problem of real-world fine-grained face retouching detection.

cs.CV↗

arXiv2023

No driver, No Regulation? --Online Legal Driving Behavior Monitoring for Self-driving Vehicles

Defined traffic laws must be respected by all vehicles. However, it is essential to know which behaviors violate the current laws, especially when a responsibility issue is involved in an accident. This brings challenges of digitizing human-driver-oriented traffic laws and monitoring vehicles' behaviors continuously. To address these challenges, this paper aims to digitize traffic law comprehensively and provide an application for online monitoring of legal driving behavior for autonomous vehicles. This paper introduces a layered trigger domain-based traffic law digitization architecture with digitization-classified discussions and detailed atomic propositions for online monitoring. The principal laws on a highway and at an intersection are taken as examples, and the corresponding logic and atomic propositions are introduced in detail. Finally, the digitized traffic laws are verified on the Chinese highway and intersection datasets, and defined thresholds are further discussed according to the driving behaviors in the considered dataset. This study can help manufacturers and the government in defining specifications and laws and can also be used as a useful reference in traffic laws compliance decision-making. Source code is available on https://github.com/SOTIF-AVLab/DOTL.

eess.SY↗