arXiv subjects

Tony Zhang

Publications and source records attributed to Tony Zhang.

9 recordsLinked to original sources

arXiv2025

A Digital Twin Framework for Metamorphic Testing of Autonomous Driving Systems Using Generative Model

Ensuring the safety of self-driving cars remains a major challenge due to the complexity and unpredictability of real-world driving environments. Traditional testing methods face significant limitations, such as the oracle problem, which makes it difficult to determine whether a system's behavior is correct, and the inability to cover the full range of scenarios an autonomous vehicle may encounter. In this paper, we introduce a digital twin-driven metamorphic testing framework that addresses these challenges by creating a virtual replica of the self-driving system and its operating environment. By combining digital twin technology with AI-based image generative models such as Stable Diffusion, our approach enables the systematic generation of realistic and diverse driving scenes. This includes variations in weather, road topology, and environmental features, all while maintaining the core semantics of the original scenario. The digital twin provides a synchronized simulation environment where changes can be tested in a controlled and repeatable manner. Within this environment, we define three metamorphic relations inspired by real-world traffic rules and vehicle behavior. We validate our framework in the Udacity self-driving simulator and demonstrate that it significantly enhances test coverage and effectiveness. Our method achieves the highest true positive rate (0.719), F1 score (0.689), and precision (0.662) compared to baseline approaches. This paper highlights the value of integrating digital twins with AI-powered scenario generation to create a scalable, automated, and high-fidelity testing solution for autonomous vehicle safety.

cs.RO↗

arXiv2025

TARDIS STRIDE: A Spatio-Temporal Road Image Dataset and World Model for Autonomy

World models aim to simulate environments and enable effective agent behavior. However, modeling real-world environments presents unique challenges as they dynamically change across both space and, crucially, time. To capture these composed dynamics, we introduce a Spatio-Temporal Road Image Dataset for Exploration (STRIDE) permuting 360-degree panoramic imagery into rich interconnected observation, state and action nodes. Leveraging this structure, we can simultaneously model the relationship between egocentric views, positional coordinates, and movement commands across both space and time. We benchmark this dataset via TARDIS, a transformer-based generative world model that integrates spatial and temporal dynamics through a unified autoregressive framework trained on STRIDE. We demonstrate robust performance across a range of agentic tasks such as controllable photorealistic image synthesis, instruction following, autonomous self-control, and state-of-the-art georeferencing. These results suggest a promising direction towards sophisticated generalist agents--capable of understanding and manipulating the spatial and temporal aspects of their material environments--with enhanced embodied reasoning capabilities. Training code, datasets, and model checkpoints are made available at https://huggingface.co/datasets/Tera-AI/STRIDE.

cs.CV↗

arXiv2025

Optically accessible high-finesse millimeter-wave resonator for cavity quantum electrodynamics with atom arrays

Cavity quantum electrodynamics (QED) is a powerful tool in quantum science, enabling preparation of non-classical states of light and scalable entanglement of many atoms coupled to a single field mode. While the most coherent atom-photon interactions have been achieved using superconducting millimeter-wave cavities coupled to Rydberg atoms, these platforms so far lack the optical access required for trapping and addressing individual atomic qubits. We present a millimeter-wave Fabry-P\'erot cavity with finesse $5.8(1) \times 10^7$ at a temperature of 1 K providing generous transverse optical access (numerical aperture 0.56). Conflicting goals of strong atom-photon coupling and optical access motivate a near-confocal geometry. Close to confocality, however, post-paraxial corrections to the cavity spectrum introduce unexpected degeneracies between transverse modes, leading to excess cavity loss. Modeling these corrections allows for tuning the cavity geometry to evade this loss, producing a high finesse that will enable cavity QED experiments with trapped atoms deep in the strong coupling regime.

quant-ph↗

arXiv2025

InhibiDistilbert: Knowledge Distillation for a ReLU and Addition-based Transformer

This work explores optimizing transformer-based language models by integrating model compression techniques with inhibitor attention, a novel alternative attention mechanism. Inhibitor attention employs Manhattan distances and ReLU activations instead of the matrix multiplications and softmax activation of the conventional scaled dot-product attention. This shift offers potential computational and energy savings while maintaining model effectiveness. We propose further adjustments to improve the inhibitor mechanism's training efficiency and evaluate its performance on the DistilBERT architecture. Our knowledge distillation experiments indicate that the modified inhibitor transformer model can achieve competitive performance on standard NLP benchmarks, including General Language Understanding Evaluation (GLUE) and sentiment analysis tasks.

cs.CL↗

arXiv2025

AI-Augmented Metamorphic Testing for Comprehensive Validation of Autonomous Vehicles

Self-driving cars have the potential to revolutionize transportation, but ensuring their safety remains a significant challenge. These systems must navigate a variety of unexpected scenarios on the road, and their complexity poses substantial difficulties for thorough testing. Conventional testing methodologies face critical limitations, including the oracle problem determining whether the systems behavior is correct and the inability to exhaustively recreate a range of situations a self-driving car may encounter. While Metamorphic Testing (MT) offers a partial solution to these challenges, its application is often limited by simplistic modifications to test scenarios. In this position paper, we propose enhancing MT by integrating AI-driven image generation tools, such as Stable Diffusion, to improve testing methodologies. These tools can generate nuanced variations of driving scenarios within the operational design domain (ODD)for example, altering weather conditions, modifying environmental elements, or adjusting lane markings while preserving the critical features necessary for system evaluation. This approach enables reproducible testing, efficient reuse of test criteria, and comprehensive evaluation of a self-driving systems performance across diverse scenarios, thereby addressing key gaps in current testing practices.

cs.SE↗

arXiv2023

Inhibitor Transformers and Gated RNNs for Torus Efficient Fully Homomorphic Encryption

This paper introduces efficient modifications to neural network-based sequence processing approaches, laying new grounds for scalable privacy-preserving machine learning under Fully Homomorphic Encryption (FHE). Transformers are now ubiquitous in AI applications and have largely supplanted Gated Recurrent Neural Networks (RNNs) as the standard architecture for sequence modeling. Both architectures rely on costly multiplications and complex activations that hinder encrypted inference. We focus on TFHE, which supports deep circuit evaluation and efficient univariate function evaluation but makes variable-to-variable multiplication particularly expensive. To address this, we propose inhibitor designs for Transformers and gated RNNs that replace multiplications and Softmax/Sigmoid activations with additive and ReLU-based operations. These changes enable integer-only computation, reduce circuit depth, and improve the efficiency of encrypted execution while preserving learning capacity. We present complexity analyses and scaling experiments that indicate significant reductions in circuit depth and execution time under TFHE, with 3-6 times speedup for encrypted inference and 30-50% reductions in plaintext inference time. Empirical evaluations on MNIST, IMDB, and IAM handwriting show inhibitor-based models maintain competitive accuracy. Knowledge distillation further demonstrates that an inhibitor-based DistilBERT achieves performance close to that of the conventional attention model on GLUE, positioning these architectures as a viable approach for scalable, privacy-preserving AI.

cs.LG↗

arXiv2021

Covariance matrices for variance-suppressed simulations

Cosmological $N$-body simulations provide numerical predictions of the structure of the Universe against which to compare data from ongoing and future surveys, but the growing volume of the Universe mapped by surveys requires correspondingly lower statistical uncertainties in simulations, usually achieved by increasing simulation sizes at the expense of computational power. It was recently proposed to reduce simulation variance without incurring additional computational costs by adopting fixed-amplitude initial conditions. This method has been demonstrated not to introduce bias in various statistics, including the two-point statistics of galaxy samples typically used for extracting cosmological parameters from galaxy redshift survey data, but requires us to revisit current methods for estimating covariance matrices of clustering statistics for simulations. In this work, we find that it is not trivial to construct covariance matrices analytically for fixed-amplitude simulations, but we demonstrate that EZmock (Effective Zel'dovich approximation mock catalogue), the most efficient method for constructing mock catalogues with accurate two- and three-point statistics, provides reasonable covariance matrix estimates for such simulations. We further examine how the variance suppression obtained by amplitude-fixing depends on three-point clustering, small-scale clustering, and galaxy bias, and propose intuitive explanations for the effects we observe based on the EZmock bias model.

astro-ph.CO↗

arXiv2021

Semi-Supervised Audio Representation Learning for Modeling Beehive Strengths

Honey bees are critical to our ecosystem and food security as a pollinator, contributing 35% of our global agriculture yield. In spite of their importance, beekeeping is exclusively dependent on human labor and experience-derived heuristics, while requiring frequent human checkups to ensure the colony is healthy, which can disrupt the colony. Increasingly, pollinator populations are declining due to threats from climate change, pests, environmental toxicity, making their management even more critical than ever before in order to ensure sustained global food security. To start addressing this pressing challenge, we developed an integrated hardware sensing system for beehive monitoring through audio and environment measurements, and a hierarchical semi-supervised deep learning model, composed of an audio modeling module and a predictor, to model the strength of beehives. The model is trained jointly on audio reconstruction and prediction losses based on human inspections, in order to model both low-level audio features and circadian temporal dynamics. We show that this model performs well despite limited labels, and can learn an audio embedding that is useful for characterizing different sound profiles of beehives. This is the first instance to our knowledge of applying audio-based deep learning to model beehives and population size in an observational setting across a large number of hives.

cs.SD↗

arXiv2021

HVAQ: A High-Resolution Vision-Based Air Quality Dataset

Air pollutants, such as particulate matter, negatively impact human health. Most existing pollution monitoring techniques use stationary sensors, which are typically sparsely deployed. However, real-world pollution distributions vary rapidly with position and the visual effects of air pollution can be used to estimate concentration, potentially at high spatial resolution. Accurate pollution monitoring requires either densely deployed conventional point sensors, at-a-distance vision-based pollution monitoring, or a combination of both. The main contribution of this paper is that to the best of our knowledge, it is the first publicly available, high temporal and spatial resolution air quality dataset containing simultaneous point sensor measurements and corresponding images. The dataset enables, for the first time, high spatial resolution evaluation of image-based air pollution estimation algorithms. It contains PM2.5, PM10, temperature, and humidity data. We evaluate several state-of-art vision-based PM concentration estimation algorithms on our dataset and quantify the increase in accuracy resulting from higher point sensor density and the use of images. It is our intent and belief that this dataset can enable advances by other research teams working on air quality estimation. Our dataset is available at https://github.com/implicitDeclaration/HVAQ-dataset/tree/master.

cs.CV↗