arXiv subjects

Liying Liu

Publications and source records attributed to Liying Liu.

4 recordsLinked to original sources

arXiv2026

Provenance Tracking in AI Compilers through the Lens of Coalgebra

AI compilers aggressively rewrite computation graphs through normalization, lowering, and optimization, making it difficult to track the provenance of tensors and operators across compilation. Reliable provenance is essential for attaching platform-specific postprocessing, debugging compiler behavior, and validating transformations, yet existing solutions are either invasive or ad hoc under non-injective graph rewrites. We present a lightweight, generative approach to provenance tracking based on observational semantics. Instead of propagating identifiers through compiler passes, we observe graph transformations and reason about provenance in terms of observable computational actions. We formalize this approach using a coalgebraic model and bisimulation, which preserves provenance even when intermediate nodes are eliminated. Furthermore, we implement this approach in a prototype AI compiler COVAN, demonstrating stable provenance across compilation pipelines with minimal engineering overhead.

cs.DB↗

arXiv2026

NTR: Neural Token Reconstruction for Scene Token Bottleneck in End-to-End Driving

Recent perception-free end-to-end (E2E) autonomous driving methods bypass explicit perception outputs by compressing dense image patch tokens into compact scene tokens for downstream trajectory generation and scoring. While these scene tokens form a compact visual bottleneck for the planner, they receive supervision solely from the planning objective, providing limited constraints on the encoded visual information. To address this limitation, we introduce Neural Token Reconstruction (NTR), a representation learning framework to directly constrain the compact scene-token bottleneck in perception-free driving. NTR introduces a self-distillation masked latent reconstruction objective that reconstructs masked patch-level latent features using only compact scene tokens as reconstruction memory. This forces reconstruction gradients to pass exclusively through the scene-token bottleneck, encouraging scene tokens to preserve richer and less redundant visual representations for planning. We further introduce semantic priors derived from foundation-model annotations as a weak semantic interface biasing reconstruction targets toward driving-related structures without introducing explicit perception heads. All auxiliary reconstruction components are removed at inference time, leaving the deployed planner unchanged. NTR achieves state-of-the-art performance on three public autonomous driving benchmarks, including 8.0461 RFS on Waymo E2E and 94.1 PDMS / 90.9 EPDMS on NavSim1&2. The learned scene tokens exhibit lower pairwise redundancy and higher effective rank, indicating that effective bottleneck supervision improves both compact visual representation learning and planning performance.

cs.CV↗

arXiv2025

Improvement of H$_2$O$_2$ electrogeneration using a Vulcan XC72 carbon-based electrocatalyst modified with Ce-doped Nb$_2$O$_5$

The use of the oxygen reduction reaction (ORR) for in-situ production of H$_2$O$_2$ is an attractive alternative to replace the methods based on anthraquinone oxidation. This study investigates the modification of Vulcan XC72 carbon with Ce-doped Nb$_2$O$_5$ in different molar proportions and its application as electrocatalysts in the ORR. One performed the characterization of the electrocatalysts using X-ray diffraction, Raman spectroscopy, scanning electron microscopy, transmission electron microscopy, contact angle measurements, and X-ray photoelectron spectroscopy. Subsequently, the electrocatalysts were analyzed for the ORR and the Nb$_2$O$_5$ doped with 0.5% Ce showing the highest electrocatalytic response. This electrocatalyst was also employed as a gas diffusion electrode and exhibited more significant H$_2$O$_2$ production at all potentials than the Vulcan XC72 carbon modified solely with Nb$_2$O$_5$. At the applied potentials of -1.3 V and -1.9 V, it produced 105% and 86% more H$_2$O$_2$, respectively, than the Vulcan XC72 carbon modified only with Nb$_2$O$_5$. These results can be attributed to the doping of Nb$_2$O$_5$ with 0.5% Ce, which induces local distortions in the crystal lattice of Nb$_2$O$_5$ due to the difference in ionic radius between Nb$^{5+}$ and Ce$^{3+}$, which combined with increased hydrophilicity and wetting properties, may have facilitated electron transfer and O$_2$ transport, favoring the ORR.

cond-mat.mtrl-sci↗

arXiv2025

Compound and Parallel Modes of Tropical Convolutional Neural Networks

Convolutional neural networks (CNNs) are foundational to many state-of-the-art computer vision systems, yet their reliance on multiplication-intensive computations poses challenges for deployment on resource-constrained devices. While tropical convolutional neural networks (TCNNs) reduce this computational burden by replacing multiplications with cheaper min/maxplus operations, they often do so at the cost of reduced model accuracy. To address this tradeoff, we introduce two novel extensions of tropical convolution: compound tropical convolution (cTCNN) and parallel tropical convolution (pTCNN). These operators combine minplus and maxplus algebraic operations within a single layer to enhance representational capacity while maintaining low computational cost. We provide an open-source implementation of these operators in a PyTorch-compatible framework, featuring optimized GPU kernels developed with TileLang. Through extensive experiments on image classification and semantic segmentation benchmarks, we demonstrate that our proposed cTCNN and pTCNN layers achieve competitive performance against standard CNNs while significantly reducing the number of multiplications. Moreover, we show that hybrid models, which integrate both tropical and conventional convolutions, can further improve the accuracy-efficiency balance. Our findings suggest that these tropical convolution variants are viable and effective components for building efficient deep learning models

cs.CV↗