arXiv subjects

Jinfeng Wei

Publications and source records attributed to Jinfeng Wei.

5 recordsLinked to original sources

arXiv2026

Measurement-Feedback Quantum Information Engine: Coherence-Transition Interference and Correlated Work Statistics

Measurement and feedback jointly prepare coherence and select finite-time dynamics in quantum information engines. Complementing a companion experimental realization, we develop a mechanism-resolved theory of the resulting work statistics and temporal correlations. The con?ditional work separates into population transfer and a phase-sensitive coherence-transition inter?ference term. Symmetric full counting statistics maps this interference to an equal-and-opposite half-quantum pair in the work quasiprobability, entering odd moments while leaving even moments fixed by endpoint mixing. An outcome-resolved tilted kernel then propagates these statistics through the correlated measurement record, yielding finite-cycle and fixed-time fluctuation corrections. The same memory reduces the reversible record-reset cost from the one-symbol entropy to the entropy rate. Our results link coherent work statistics, feedback memory, and information thermodynamics

quant-ph↗

arXiv2026

Hidden APIs in Language Models: Discovering Reusable Causal Interfaces from Forked Futures

Identical language-model answers can arise from hidden states that support different future computations, so current-answer probes do not establish a reusable internal interface. We introduce forked futures: future operations are sampled only after a prefix state has formed, and states are compared through the response distributions induced by those operations. This yields an empirical causal quotient over hidden states without requiring researcher-specified latent labels. Shared, Local, Mixture, and Distributed interfaces then compete under prequential causal description length subject to future-signature fidelity and matched capacity constraints. In the two detailed model evaluations, Shared has the lowest held-out description length, with gains of 0.216 nats on Qwen2.5-1.5B and 0.294 nats on Llama-3-8B, while maintaining tightly clustered mean future-signature distortion; a five-backbone sweep preserves the positive direction of Sharedness Gain. The figure-aligned transplantation analysis gives Shared the strongest joint target-correctness, locality, copy-preservation, and composite profile, and API-aligned paths mediate 0.749 of the target effect versus 0.150 for matched null paths. In the blind four-class model-organism test, 14/16 architectures are recovered, with one observed non-Shared to Shared error among 12 non-Shared organisms. These results support an economical reusable causal interface within the tested operation banks, while keeping the claim explicitly conditional on the candidate architectures, interventions, and held-out futures.

cs.AI↗

arXiv2026

Experimental Demonstration of a Measurement-Feedback Quantum Information Engine

Harnessing finite-time nonadiabatic transitions that are conventionally associated with quantum inner friction for useful work extraction remains an open experimental challenge. Here we address this issue by introducing and experimentally realizing an innovative measurement-feedback quantum information engine model in the trapped 40Ca+ ion system, in which the projective measurement replaces the hot reservoir as a nonthermal energy source and the feedback control conditionally steers the system through either unitary compression-expansion strokes or thermalization. We experimentally show that the repeated feedback cycle converges to a stable operating regime with a resolved energetic balance, and, by controlling the measurement angle and stroke duration, measurement-induced coherence and the finite-time nonadiabatic contribution can enhance work extraction and raise the efficiency above the corresponding Otto benchmark. The experimental results further show that, over finite ranges of stroke durations, the efficiency and intrinsic cycle power can increase simultaneously. Our experiment establishes a route toward information-to-work quantum engines that convert finite-time irreversibility into performance-enhancing resources.

quant-ph↗

arXiv2024

DOPRA: Decoding Over-accumulation Penalization and Re-allocation in Specific Weighting Layer

In this work, we introduce DOPRA, a novel approach designed to mitigate hallucinations in multi-modal large language models (MLLMs). Unlike existing solutions that typically involve costly supplementary training data or the integration of external knowledge sources, DOPRA innovatively addresses hallucinations by decoding specific weighted layer penalties and redistribution, offering an economical and effective solution without additional resources. DOPRA is grounded in unique insights into the intrinsic mechanisms controlling hallucinations within MLLMs, especially the models' tendency to over-rely on a subset of summary tokens in the self-attention matrix, neglecting critical image-related information. This phenomenon is particularly pronounced in certain strata. To counteract this over-reliance, DOPRA employs a strategy of weighted overlay penalties and redistribution in specific layers, such as the 12th layer, during the decoding process. Furthermore, DOPRA includes a retrospective allocation process that re-examines the sequence of generated tokens, allowing the algorithm to reallocate token selection to better align with the actual image content, thereby reducing the incidence of hallucinatory descriptions in auto-generated captions. Overall, DOPRA represents a significant step forward in improving the output quality of MLLMs by systematically reducing hallucinations through targeted adjustments during the decoding process.

cs.CL↗

arXiv2021

Dynamic Hypergraph Convolutional Networks for Skeleton-Based Action Recognition

Graph convolutional networks (GCNs) based methods have achieved advanced performance on skeleton-based action recognition task. However, the skeleton graph cannot fully represent the motion information contained in skeleton data. In addition, the topology of the skeleton graph in the GCN-based methods is manually set according to natural connections, and it is fixed for all samples, which cannot well adapt to different situations. In this work, we propose a novel dynamic hypergraph convolutional networks (DHGCN) for skeleton-based action recognition. DHGCN uses hypergraph to represent the skeleton structure to effectively exploit the motion information contained in human joints. Each joint in the skeleton hypergraph is dynamically assigned the corresponding weight according to its moving, and the hypergraph topology in our model can be dynamically adjusted to different samples according to the relationship between the joints. Experimental results demonstrate that the performance of our model achieves competitive performance on three datasets: Kinetics-Skeleton 400, NTU RGB+D 60, and NTU RGB+D 120.

cs.CV↗