Selected work

Representative Papers

Motion Estimation for Multi-Object Tracking using KalmanNet with Semantic-Independent Encoding

Sep 14, 2025

To address the degraded estimation performance of linear constant-velocity Kalman filtering in multi-object tracking—caused by model mismatch and non-stationary motion—this paper proposes a learning-enhanced Semantic-Independent Kalman Filter (SIC-KF). Our core innovation is a lightweight semantic-independent encoder comprising 1D convolutions (kernel size 1), fully connected layers, and nonlinear activation modules, which jointly learns decoupled representations of homogeneous semantic features and captures nonlinear dependencies among heterogeneous elements, thereby enhancing motion-awareness of the state vector. SIC-KF is seamlessly embedded into the KalmanNet framework to enable end-to-end differentiable training. Experiments on a large-scale, semi-synthetic dataset constructed by us demonstrate that SIC-KF significantly outperforms conventional Kalman filters and state-of-the-art learning-based filters, achieving new SOTA performance in both trajectory prediction accuracy and robustness to motion abruptness.

0 citationsRead paper

Marker Gene Method : Identifying Stable Solutions in a Dynamic Environment

Jun 30, 2025

Competitive coevolutionary algorithms (CCEAs) suffer from unstable convergence in dynamic environments due to non-transitivity and the Red Queen effect. Method: This paper proposes a *tag-gene* approach: (i) employing evolvable tag genes as dynamic performance benchmarks; (ii) integrating an adaptive weighting mechanism to balance exploration and exploitation; and (iii) incorporating a memory pool to enhance utilization of historical information. Contribution/Results: We theoretically prove that the method establishes a strong attractor under strict competitive games, with convergence points asymptotically approaching Nash equilibria. Empirical evaluation across canonical dynamic benchmarks—including Rock-Paper-Scissors, ZDT multi-objective optimization, and Shapley-biased games—demonstrates significant improvements in stability, robustness, and convergence quality. The proposed framework provides a novel, interpretable, and theoretically grounded paradigm for dynamic coevolution.

0 citationsRead paper

Recent publications

Latest Papers

Motion Estimation for Multi-Object Tracking using KalmanNet with Semantic-Independent Encoding

Sep 14, 2025

To address the degraded estimation performance of linear constant-velocity Kalman filtering in multi-object tracking—caused by model mismatch and non-stationary motion—this paper proposes a learning-enhanced Semantic-Independent Kalman Filter (SIC-KF). Our core innovation is a lightweight semantic-independent encoder comprising 1D convolutions (kernel size 1), fully connected layers, and nonlinear activation modules, which jointly learns decoupled representations of homogeneous semantic features and captures nonlinear dependencies among heterogeneous elements, thereby enhancing motion-awareness of the state vector. SIC-KF is seamlessly embedded into the KalmanNet framework to enable end-to-end differentiable training. Experiments on a large-scale, semi-synthetic dataset constructed by us demonstrate that SIC-KF significantly outperforms conventional Kalman filters and state-of-the-art learning-based filters, achieving new SOTA performance in both trajectory prediction accuracy and robustness to motion abruptness.

0 citationsRead paper

Marker Gene Method : Identifying Stable Solutions in a Dynamic Environment

Jun 30, 2025

Competitive coevolutionary algorithms (CCEAs) suffer from unstable convergence in dynamic environments due to non-transitivity and the Red Queen effect. Method: This paper proposes a *tag-gene* approach: (i) employing evolvable tag genes as dynamic performance benchmarks; (ii) integrating an adaptive weighting mechanism to balance exploration and exploitation; and (iii) incorporating a memory pool to enhance utilization of historical information. Contribution/Results: We theoretically prove that the method establishes a strong attractor under strict competitive games, with convergence points asymptotically approaching Nash equilibria. Empirical evaluation across canonical dynamic benchmarks—including Rock-Paper-Scissors, ZDT multi-objective optimization, and Shapley-biased games—demonstrates significant improvements in stability, robustness, and convergence quality. The proposed framework provides a novel, interpretable, and theoretically grounded paradigm for dynamic coevolution.

0 citationsRead paper