A Computationally Feasible Framework for Causal Probabilistic Explanation
本文提出概率因果影响(PCI)框架，通过蒙特卡洛方法估计因果模型，解决了大规模模型中因果解释的计算难题。
0 citationsRead paper
Selected work
Representative Papers
Wasserstein Residuals: Learning Gradient Flows from Population Dynamics
This work addresses the reconstruction of population dynamics governed by Wasserstein gradient flows from sparse observational data. The authors propose a particle-based "stitching" method that bypasses the traditional Jordan–Kinderlehrer–Otto (JKO) time discretization, instead enforcing the continuity equation via a non-negative residual loss and integrating a data-fidelity divergence into a unified optimization objective. By eliminating the need for costly optimal transport computations and fixed time steps, the approach is simulation-agnostic and robust to irregular or widely spaced observation intervals. Evaluated on multiple trajectory inference benchmarks, the method achieves state-of-the-art performance, demonstrating particular superiority in regimes with highly sparse observations or large temporal gaps between measurements.
0 citationsRead paper
ESC: Emotional Self-Correction for Reliable Vision-Language Models
This work addresses the unreliability of existing vision-language models in multimodal reasoning, a limitation exacerbated by current self-correction approaches that rely on additional training or complex feedback mechanisms with high computational overhead. To overcome this, the paper proposes ESC—an entirely training-free self-correction framework that introduces emotional signals as a novel triggering mechanism. ESC employs an external validator to detect erroneous responses and injects affective feedback to prompt the model to autonomously reflect on and revise its outputs. Without modifying the model architecture or requiring any extra training, ESC significantly enhances model reliability across multiple benchmarks, including safety, hallucination suppression, visual perception, and multimodal reasoning, while preserving overall performance.
0 citationsRead paper
Dynestyx: A Probabilistic Programming Library for Dynamical Systems
Existing probabilistic programming languages lack native support for dynamic systems—particularly state-space models—hindering the broader adoption of Bayesian methods in this domain. This work introduces dynestyx, a library that provides first-class, unified, and user-friendly support for state-space models within a probabilistic programming framework. dynestyx enables flexible specification of priors, accommodates both discrete- and continuous-time dynamics, handles mixed-effects data, and facilitates joint Bayesian inference over latent states and model parameters with full uncertainty quantification. By doing so, this contribution substantially enhances the accessibility, flexibility, and practical utility of dynamic system modeling across statistics, signal processing, and machine learning.
0 citationsRead paper
Domain Generalization through Spatial Relation Induction over Visual Primitives
This work addresses the performance bottleneck in domain generalization caused by neglecting structured compositional representations. It proposes a novel approach that explicitly models visual primitives and their higher-order spatial relationships through differentiable binary-to-quaternary predicates, which capture spatial alignments among primitives. The method jointly learns primitive representations and relational structures in an end-to-end relation-induced learning framework. Specifically, a CNN backbone generates primitive heatmaps, which are processed through a concept bottleneck layer and a structural scoring layer to compute classification probabilities based on class-specific relational compositions. Evaluated on CUB-DG, the model achieves over a 4.5 percentage point improvement in accuracy, and matches state-of-the-art performance on the DomainBed benchmark, marking the first explicit incorporation of higher-order spatial relations into domain generalization.
0 citationsRead paper
Recent publications
Latest Papers
A Computationally Feasible Framework for Causal Probabilistic Explanation
本文提出概率因果影响(PCI)框架，通过蒙特卡洛方法估计因果模型，解决了大规模模型中因果解释的计算难题。
0 citationsRead paper
Wasserstein Residuals: Learning Gradient Flows from Population Dynamics
This work addresses the reconstruction of population dynamics governed by Wasserstein gradient flows from sparse observational data. The authors propose a particle-based "stitching" method that bypasses the traditional Jordan–Kinderlehrer–Otto (JKO) time discretization, instead enforcing the continuity equation via a non-negative residual loss and integrating a data-fidelity divergence into a unified optimization objective. By eliminating the need for costly optimal transport computations and fixed time steps, the approach is simulation-agnostic and robust to irregular or widely spaced observation intervals. Evaluated on multiple trajectory inference benchmarks, the method achieves state-of-the-art performance, demonstrating particular superiority in regimes with highly sparse observations or large temporal gaps between measurements.
0 citationsRead paper
ESC: Emotional Self-Correction for Reliable Vision-Language Models
This work addresses the unreliability of existing vision-language models in multimodal reasoning, a limitation exacerbated by current self-correction approaches that rely on additional training or complex feedback mechanisms with high computational overhead. To overcome this, the paper proposes ESC—an entirely training-free self-correction framework that introduces emotional signals as a novel triggering mechanism. ESC employs an external validator to detect erroneous responses and injects affective feedback to prompt the model to autonomously reflect on and revise its outputs. Without modifying the model architecture or requiring any extra training, ESC significantly enhances model reliability across multiple benchmarks, including safety, hallucination suppression, visual perception, and multimodal reasoning, while preserving overall performance.
0 citationsRead paper
Dynestyx: A Probabilistic Programming Library for Dynamical Systems
Existing probabilistic programming languages lack native support for dynamic systems—particularly state-space models—hindering the broader adoption of Bayesian methods in this domain. This work introduces dynestyx, a library that provides first-class, unified, and user-friendly support for state-space models within a probabilistic programming framework. dynestyx enables flexible specification of priors, accommodates both discrete- and continuous-time dynamics, handles mixed-effects data, and facilitates joint Bayesian inference over latent states and model parameters with full uncertainty quantification. By doing so, this contribution substantially enhances the accessibility, flexibility, and practical utility of dynamic system modeling across statistics, signal processing, and machine learning.
0 citationsRead paper
Domain Generalization through Spatial Relation Induction over Visual Primitives
This work addresses the performance bottleneck in domain generalization caused by neglecting structured compositional representations. It proposes a novel approach that explicitly models visual primitives and their higher-order spatial relationships through differentiable binary-to-quaternary predicates, which capture spatial alignments among primitives. The method jointly learns primitive representations and relational structures in an end-to-end relation-induced learning framework. Specifically, a CNN backbone generates primitive heatmaps, which are processed through a concept bottleneck layer and a structural scoring layer to compute classification probabilities based on class-specific relational compositions. Evaluated on CUB-DG, the model achieves over a 4.5 percentage point improvement in accuracy, and matches state-of-the-art performance on the DomainBed benchmark, marking the first explicit incorporation of higher-order spatial relations into domain generalization.
0 citationsRead paper