SearcharxivSearch

arXiv subjects

Ziwei Cheng

Publications and source records attributed to Ziwei Cheng.

2 recordsLinked to original sources

Energy Transport Structure and Fluctuation Theorem in Nonreciprocal Harmonic Chains

Nonreciprocal interactions fundamentally alter energy transport by breaking the symmetry between forward and backward responses. Here, we uncover an exact transport structure for a one-dimensional harmonic chain with asymmetric nearest-neighbor couplings. Using a Green-function approach, we demonstrate that the direction of heat transport is determined not solely by the temperature bias, but by its competition with an effective directional asymmetry. This competition can reverse the direction of heat flow, enabling cold-to-hot transport. We further show that nonreciprocity introduces an additional power channel associated with the antisymmetric sector of the interaction, leading to a generalized steady-state energy balance involving two reservoir heat currents and a nonreciprocal power current. At the fluctuation level, the heat exchanges with the two reservoirs constitute correlated yet distinct stochastic currents, whose joint scaled cumulant generating function obeys an exact Gallavotti-Cohen symmetry. Together, these results establish a unified energetic and fluctuation framework for nonreciprocal heat transport and demonstrate that directional interactions can serve as an active resource for controlling nonequilibrium energy flows.

cond-mat.stat-mech

DoubleHelix: Structured Cross-Modal Fusion for Audio-Visual Speech Recognition with LLMs

Audio-visual speech recognition (AVSR) relies on effective fusion of audio and visual modalities, yet existing approaches treat cross-modal interaction as a single-step operation without structured iterative refinement. We present DoubleHelix, a multimodal fusion framework that reformulates fusion as an iterative cross-modal interaction process with adaptive degradation-aware enhancement. The framework comprises three components including ReverseParallelHelix for multi-turn structured interaction with learned alignment constraints, QualitySensor for learning degradation-aware gating signals, and HelixReplication for consistency-guided conditional feature enhancement. Experiments on LRS3 demonstrate that DoubleHelix achieves 0.68% WER on clean audio, outperforming previous best results by 5.6% relative improvement under matched backbone settings. Comprehensive ablation studies validate each component contribution, including targeted analysis of design choices such as asymmetric pathway weighting. The framework shows improved robustness under evaluated babble-noise conditions, achieving 11.6% WER at SNR -5dB.

cs.SD