arXiv subjects

Xinlin Wang

Publications and source records attributed to Xinlin Wang.

10 recordsLinked to original sources

arXiv2026

WA-JEPA: Rethinking the Video JEPA Paradigm for World-Action Modeling in Autonomous Driving

Video Joint Embedding Predictive Architecture (V-JEPA) learns powerful spatiotemporal representations from video through self-supervised latent feature prediction. However, V-JEPA is built around random-mask completion and deterministic regression, making it fundamentally ill-suited for autonomous driving planning that demands future-directed prediction tightly coupled with action. To address this, we rethink the V-JEPA paradigm and present WA-JEPA, a V-JEPA-native world-action model designed for autonomous driving planning. Instead of random spatiotemporal masking, WA-JEPA employs hybrid future-masked pre-training, where the model infers future latents from observed context. Departing from deterministic regression, we recast future prediction as conditional flow matching over latent futures, which substantially improves the model's ability to generate plausible future latents for downstream planning. Finally, a joint future-action predictor is proposed to denoise future scene tokens and ego trajectories together in a unified spatiotemporal latent space, allowing action supervision to directly shape planning-relevant world representations. Pre-trained on nuPlan videos and fine-tuned on NAVSIM, WA-JEPA reaches 91.7 EPDMS on NAVSIM-v2, surpassing the strongest end-to-end and world-action baselines by 1.6 and 1.3 EPDMS, and, without HUGSIM-specific fine-tuning, attains the best HD-Score of 0.4462 on the closed-loop HUGSIM benchmark under the same evaluation protocol. These results validate V-JEPA-native world-action modeling as a powerful and scalable paradigm for autonomous driving planning. Code is available at https://github.com/AFARI-Research/WA-JEPA.

cs.CV↗

arXiv2026

ChainFlow-VLA: Causal Flow Planning with Vision-Language Models

Current end-to-end autonomous driving systems are fundamentally limited by a mismatch between temporal causal reasoning and global trajectory consistency. Autoregressive (AR) models capture interaction-aware temporal dependencies via causal factorization, but their step-wise decoding leads to error accumulation and suboptimal global structure. In contrast, diffusion models optimize trajectories globally but lack explicit causal constraints, making them unreliable in interactive and safety-critical scenarios. This dichotomy reveals a deeper issue: existing methods treat causal modeling and global optimization as separate paradigms, without a principled way to unify them within a single trajectory distribution. To address this, we propose ChainFlow-VLA, which unifies causal generation and global refinement within a unified probabilistic framework. We formulate planning as a mixture over AR-induced modes and learn Vision-Language Model (VLM)-conditioned residual distributions over these modes. An autoregressive generator (Chain) produces a discrete set of causal trajectory modes, followed by a diffusion-based refiner (Flow) that leverages VLM hidden states as semantic priors to perform mode-conditioned correction in residual space while preserving causal structure. This straightforward conditioning seamlessly injects high-level scene understanding into fine-grained trajectory adjustments. Experiments demonstrate that ChainFlow-VLA achieves robust planning in ambiguous and long-tail scenarios, achieving a state-of-the-art score of 94.85 on the NAVSIM v1 leaderboard, matching human-level performance (94.8). Code will be available at https://github.com/AFARI-Research/ChainFlow-VLA.

cs.CV↗

arXiv2026

Rethinking Scale: Deployment Trade-offs of Small Language Models under Agent Paradigms

Despite the impressive capabilities of large language models, their substantial computational costs, latency, and privacy risks hinder their widespread deployment in real-world applications. Small Language Models (SLMs) with fewer than 10 billion parameters present a promising alternative; however, their inherent limitations in knowledge and reasoning curtail their effectiveness. Existing research primarily focuses on enhancing SLMs through scaling laws or fine-tuning strategies while overlooking the potential of using agent paradigms, such as tool use and multi-agent collaboration, to systematically compensate for the inherent weaknesses of small models. To address this gap, this paper presents the first large-scale, comprehensive study of <10B open-source models under three paradigms: (1) the base model, (2) a single agent equipped with tools, and (3) a multi-agent system with collaborative capabilities. Our results show that single-agent systems achieve the best balance between performance and cost, while multi-agent setups add overhead with limited gains. Our findings highlight the importance of agent-centric design for efficient and trustworthy deployment in resource-constrained settings.

cs.CL↗

arXiv2026

Low-Complexity Algorithm for Stackelberg Prediction Games with Global Optimality

Stackelberg prediction games (SPGs) model strategic data manipulation in adversarial learning via a leader--follower interaction between a learner and a self-interested data provider, leading to challenging bilevel optimization problems. Focusing on the least-squares setting (SPG-LS), recent work shows that the bilevel program admits an equivalent spherically constrained least-squares (SCLS) reformulation, which avoids costly conic programming and enables scalable algorithms. In this paper, we develop a simple and efficient alternating direction method of multiplier (ADMM) based solver for the SCLS problem. By introducing a consensus splitting that separates the quadratic objective from the spherical constraint, we obtain an augmented Lagrangian formulation with closed-form updates: the primal quadratic step reduces to solving a fixed shifted linear system, the constraint step is a projection onto the unit sphere, and the dual step is a lightweight scaled ascent. The resulting method has low per-iteration complexity and allows pre-factorization of the constant system matrix for substantial speedups. Experiments demonstrate that the proposed ADMM approach achieves competitive solution quality with significantly improved computational efficiency compared with existing global solvers for SCLS, particularly in sparse and high-dimensional regimes.

eess.SP↗

arXiv2025

Which Company Adjustment Matter? Insights from Uplift Modeling on Financial Health

Uplift modeling has achieved significant success in various fields, particularly in online marketing. It is a method that primarily utilizes machine learning and deep learning to estimate individual treatment effects. This paper we apply uplift modeling to analyze the effect of company adjustment on their financial status, and we treat these adjustment as treatments or interventions in this study. Although there have been extensive studies and application regarding binary treatments, multiple treatments, and continuous treatments, company adjustment are often more complex than these scenarios, as they constitute a series of multiple time-dependent actions. The effect estimation of company adjustment needs to take into account not only individual treatment traits but also the temporal order of this series of treatments. This study collects a real-world data set about company financial statements and reported behavior in Luxembourg for the experiments. First, we use two meta-learners and three other well-known uplift models to analyze different company adjustment by simplifying the adjustment as binary treatments. Furthermore, we propose a new uplift modeling framework (MTDnet) to address the time-dependent nature of these adjustment, and the experimental result shows the necessity of considering the timing of these adjustment.

cs.CE↗

arXiv2024

Datasets for Advanced Bankruptcy Prediction: A survey and Taxonomy

Bankruptcy prediction is an important research area that heavily relies on data science. It aims to help investors, managers, and regulators better understand the operational status of corporations and predict potential financial risks in advance. To improve prediction, researchers and practitioners have begun to utilize a variety of different types of data, ranging from traditional financial indicators to unstructured data, to aid in the construction and optimization of bankruptcy forecasting models. Over time, not only instrumentalized data improved, but also instrumentalized methodology for data structuring, cleaning, and analysis. With the aid of advanced analytical techniques that deploy machine learning and deep learning algorithms, bankruptcy assessment became more accurate over time. However, due to the sensitivity of financial data, the scarcity of valid public datasets remains a key bottleneck for the rapid modeling and evaluation of machine learning algorithms for targeted tasks. This study therefore introduces a taxonomy of datasets for bankruptcy research, and summarizes their characteristics. This paper also proposes a set of metrics to measure the quality and the informativeness of public datasets The taxonomy, coupled with the informativeness measure, thus aims at providing valuable insights to better assist researchers and practitioners in developing potential applications for various aspects of credit assessment and decision making by pointing at appropriate datasets for their studies.

cs.CE↗

arXiv2024

Augmenting Bankruptcy Prediction using Reported Behavior of Corporate Restructuring

Credit risk assessment of a company is commonly conducted by utilizing financial ratios that are derived from its financial statements. However, this approach may not fully encompass other significant aspects of a company. We propose the utilization of a hybrid dataset that combines financial statements with information about corporate restructuring behavior in order to construct diverse machine learning models to predict bankruptcy. Utilizing a hybrid data set provides a more comprehensive and holistic perspective on a company's financial position and the dynamics of its business operations. The experiments were carried out using publicly available records of all the files submitted by small and medium-sized enterprises to Luxembourg Business Registers. We conduct a comparative analysis of bankruptcy prediction using six machine learning models. Furthermore, we validate the effectiveness of the hybrid dataset. In addition to the conventional testing set, we deliberately chose the timeframe encompassing the years of the Covid-19 pandemic as an additional testing set in order to evaluate the robustness of the models. The experimental results demonstrate that the hybrid data set can improve the performance of the model by 4%-13% compared to a single source data set. We also identify suitable models for predicting bankruptcy.

cs.CE↗

arXiv2019

Complex Scene Classification of PolSAR Imagery based on a Self-paced Learning Approach

Existing polarimetric synthetic aperture radar (PolSAR) image classification methods cannot achieve satisfactory performance on complex scenes characterized by several types of land cover with significant levels of noise or similar scattering properties across land cover types. Hence, we propose a supervised classification method aimed at constructing a classifier based on self-paced learning (SPL). SPL has been demonstrated to be effective at dealing with complex data while providing classifier. In this paper, a novel Support Vector Machine (SVM) algorithm based on SPL with neighborhood constraints (SVM_SPLNC) is proposed. The proposed method leverages the easiest samples first to obtain an initial parameter vector. Then, more complex samples are gradually incorporated to update the parameter vector iteratively. Moreover, neighborhood constraints are introduced during the training process to further improve performance. Experimental results on three real PolSAR images show that the proposed method performs well on complex scenes.

cs.CV↗

arXiv2008

Isolated sub-100-attosecond pulse generation via controlling electron dynamics

A new method to coherently control the electron dynamics is proposed using a few-cycle laser pulse in combination with a controlling field. It is shown that this method not only broadens the attosecond pulse bandwidth, but also reduces the chirp, then an isolated 80-as pulse is straightforwardly obtained and even shorter pulse is achievable by increasing the intensity of the controlling field. Such ultrashort pulses allow one to investigate ultrafast electronic processes which have never be achieved before. In addition, the few-cycle synthesized pulse is expected to manipulate a wide range of laser-atom interactions.

physics.atom-ph↗

arXiv2007

Efficient generation of an isolated single-cycle attosecond pulse

A new method for efficiently generating an isolated single-cycle attosecond pulse is proposed. It is shown that the ultraviolet (UV) attosecond pulse can be utilized as a robust tool to control the dynamics of electron wave packets (EWPs). By adding a UV attosecond pulse to an infrared (IR) few-cycle pulse at a proper time, only one return of the EWP to the parent ion is selected to effectively contribute to the harmonics, then an isolated two-cycle 130-as pulse with a bandwidth of 45 eV is obtained. After complementing the chirp, an isolated single-cycle attosecond pulse with a duration less than 100 as seems achievable. In addition, the contribution of the quantum trajectories can be selected by adjusting the delay between the IR and UV fields. Using this method, the harmonic and attosecond pulse yields are efficiently enhanced in contrast to the scheme [G. Sansone {\it et al.}, Science {\bf314}, 443 (2006)] using a few-cycle IR pulse in combination with the polarization gating technique.

physics.atom-ph↗