arXiv subjects

Chenguang Zhang

Publications and source records attributed to Chenguang Zhang.

14 recordsLinked to original sources

arXiv2026

Isolation-aware Scheduling Framework for DNN-based End-to-End Autonomous Driving System on Tile-based Accelerators

Level-4+ autonomous driving systems (ADS) must run dozens of heterogeneous deep neural networks (DNNs) as end-to-end (E2E) pipelines under a strict latency constraint (<=100 ms), even as execution time varies by up to 3.3x. Cost rules out dedicating isolated hardware to each function in mass-produced ADS, so these DNNs must be densely colocated on a single chip, which introduces shared-resource contention. Tile-based accelerators expose two scheduling opportunities that conventional ADS schedulers do not exploit. First, they provide a tunable degree of parallelism (DoP): assigning more tiles raises DoP and can shorten DNN execution time. Second, they provide hardware-native isolation: tiles can be physically partitioned among co-located DNNs. But using this flexibility is expensive: changing a task's DoP triggers a stop-migrate-restart reallocation of its weights and intermediate features. At ADS task rates of 10-240 Hz, these stalls accumulate along E2E chains and threaten deadlines. Reservation-based schedulers fix DoP and leave this flexibility unused; work-conserving schedulers exploit it but assume reallocation is cheap and treat deadlines as independent. We present ADS-Tile that combines configurable isolation and elastic reservation into a spatio-temporal isolation-sharing space that bounds where and when reallocation occurs; a probabilistic latency model and a DAG-aware runtime scheduler then use this space to decide task colocation and DoP under shared E2E deadlines. On an industry- and academia- derived ADS benchmark, ADS-Tile uses up to 32% fewer tiles than the work-conserving baseline in deadline-critical settings and cuts reallocation-induced wasted processing capacity from 17%-44% to below 1.2%. Controlled spatio-temporal sharing improves resource efficiency and latency predictability for tile-based ADS.

cs.AR↗

arXiv2026

Traffic Sign Recognition in Autonomous Driving: Dataset, Benchmark, and Field Experiment

Traffic Sign Recognition (TSR) is a core perception capability for autonomous driving, where robustness to cross-region variation, long-tailed categories, and semantic ambiguity is essential for reliable real-world deployment. Despite steady progress in recognition accuracy, existing traffic sign datasets and benchmarks offer limited diagnostic insight into how different modeling paradigms behave under these practical challenges. We present TS-1M, a large-scale and globally diverse traffic sign dataset comprising over one million real-world images across 454 standardized categories, together with a diagnostic benchmark designed to analyze model capability boundaries. Beyond standard train-test evaluation, we provide a suite of challenge-oriented settings, including cross-region recognition, rare-class identification, low-clarity robustness, and semantic text understanding, enabling systematic and fine-grained assessment of modern TSR models. Using TS-1M, we conduct a unified benchmark across three representative learning paradigms: classical supervised models, self-supervised pretrained models, and multimodal vision-language models (VLMs). Our analysis reveals consistent paradigm-dependent behaviors, showing that semantic alignment is a key factor for cross-region generalization and rare-category recognition, while purely visual models remain sensitive to appearance shift and data imbalance. Finally, we validate the practical relevance of TS-1M through real-scene autonomous driving experiments, where traffic sign recognition is integrated with semantic reasoning and spatial localization to support map-level decision constraints. Overall, TS-1M establishes a reference-level diagnostic benchmark for TSR and provides principled insights into robust and semantic-aware traffic sign perception. Project page: https://guoyangzhao.github.io/projects/ts1m.

cs.CV↗

arXiv2025

AIM: Software and Hardware Co-design for Architecture-level IR-drop Mitigation in High-performance PIM

SRAM Processing-in-Memory (PIM) has emerged as the most promising implementation for high-performance PIM, delivering superior computing density, energy efficiency, and computational precision. However, the pursuit of higher performance necessitates more complex circuit designs and increased operating frequencies, which exacerbate IR-drop issues. Severe IR-drop can significantly degrade chip performance and even threaten reliability. Conventional circuit-level IR-drop mitigation methods, such as back-end optimizations, are resource-intensive and often compromise power, performance, and area (PPA). To address these challenges, we propose AIM, comprehensive software and hardware co-design for architecture-level IR-drop mitigation in high-performance PIM. Initially, leveraging the bit-serial and in-situ dataflow processing properties of PIM, we introduce Rtog and HR, which establish a direct correlation between PIM workloads and IR-drop. Building on this foundation, we propose LHR and WDS, enabling extensive exploration of architecture-level IR-drop mitigation while maintaining computational accuracy through software optimization. Subsequently, we develop IR-Booster, a dynamic adjustment mechanism that integrates software-level HR information with hardware-based IR-drop monitoring to adapt the V-f pairs of the PIM macro, achieving enhanced energy efficiency and performance. Finally, we propose the HR-aware task mapping method, bridging software and hardware designs to achieve optimal improvement. Post-layout simulation results on a 7nm 256-TOPS PIM chip demonstrate that AIM achieves up to 69.2% IR-drop mitigation, resulting in 2.29x energy efficiency improvement and 1.152x speedup.

cs.AR↗

arXiv2025

Data-Driven Design-Test-Make-Analyze Paradigm for Inorganic Crystals: Ultrafast Synthesis of Ternary Oxides

Data-driven methodologies hold the promise of revolutionizing inorganic materials discovery, but they often face challenges due to discrepancies between theoretical predictions and experimental validation. In this work, we present an end-to-end discovery framework that leverages synthesizability, oxidation state probability, and reaction pathway calculations to guide the exploration of transition metal oxide spaces. Two previously unsynthesized target compositions, ZnVO3 and YMoO3, passed preliminary computational evaluation and were considered for ultrafast synthesis. Comprehensive structural and compositional analysis confirmed the successful synthesis ZnVO3 in a partially disordered spinel structure, validated via Density Functional Theory (DFT). Exploration of YMoO3 led to YMoO3-x with elemental composition close to 1:1:3; the structure was subsequently identified to be Y4Mo4O11 through micro-electron diffraction (microED) analysis. Our framework effectively integrates multi-aspect physics-based filtration with in-depth characterization, demonstrating the feasibility of designing, testing, synthesizing, and analyzing (DTMA) novel material candidates, marking a significant advancement towards inorganic materials by design.

cond-mat.mtrl-sci↗

arXiv2025

Direct Joule-Heated Non-Equilibrium Synthesis Enables High Performing Thermoelectrics

High-throughput synthesis of bulk inorganic materials is crucial for accelerating functional materials discovery but is hindered by slow, energy-intensive solid-state methods. We introduce Direct Joule-Heated Synthesis (DJS), a rapid, single-step and scalable solid-state synthesis technique achieving a $10^5$-fold speedup and 20,000x energy efficiency improvement over conventional synthesis. DJS enables the synthesis of dense, bulk chalcogenides ($\mathrm{Bi_{0.5}Sb_{1.5}Te_3}$, $\mathrm{AgSbTe_2}$), achieving a zT of 2.3 at 573 K in optimally Cd/Se co-doped $\mathrm{AgSbTe_2}$, one of the highest for polycrystalline materials at this temperature. DJS enables optimal co-doping and rapid, non-equilibrium solidification, producing lamellar microstructures, interfacial regions, and cation-ordered nanodomains that scatter all-scale phonons, achieving ultralow lattice thermal conductivity (~0.2 $W m^{-1} K^{-1}$ at 573 K). DJS establishes a new benchmark for scalable and fast synthesis, accelerating functional material discovery.

cond-mat.mtrl-sci↗

arXiv2024

Information upper bounds in composite quantum systems

The intrinsic information of quantum systems refers to the information required to define a quantum state, and may reveal how the nature stores and processes microscopic information. However, there is an evident paradox due to the "\textit{information scale contrast}": Existing analytical results on the information bounds of quantum systems show that the information-carrying capacity of an $n$-qubit system is only $O(n)$. Nonetheless, the intrinsic information content (estimated by, e.g., the number of parameters or the complexity of ontic embedding) indicates that defining an $n$-qubit system often requires information of the order $O(2^n)$. In this paper, we aim to clarify the upper bound of intrinsic information in quantum systems, as well as explain and resolve the aforementioned paradox. Starting with an analysis of the dependence between the Bloch parameters, we take the structural constraints of the quantum state space as a prior and derive the posterior information content of a quantum state through the process of Maximum A Posteriori (MAP) estimation. We analytically prove that the upper bound of the posterior information content of a 2-qubit system is exactly equal to 2. Furthermore, we conjecturally generalized this result to $n$-qubit systems via a convex optimization process and numerical experiments. In summary, our main theoretical observation is that the tight structural constraints among the parameters of a quantum state make them highly interdependent, and thus unable to freely encode information. It turns out that the intrinsic information of an $n$-qubit system defined by posterior information is bounded by $n$ classical bits.

quant-ph↗

arXiv2024

TSCLIP: Robust CLIP Fine-Tuning for Worldwide Cross-Regional Traffic Sign Recognition

Traffic sign is a critical map feature for navigation and traffic control. Nevertheless, current methods for traffic sign recognition rely on traditional deep learning models, which typically suffer from significant performance degradation considering the variations in data distribution across different regions. In this paper, we propose TSCLIP, a robust fine-tuning approach with the contrastive language-image pre-training (CLIP) model for worldwide cross-regional traffic sign recognition. We first curate a cross-regional traffic sign benchmark dataset by combining data from ten different sources. Then, we propose a prompt engineering scheme tailored to the characteristics of traffic signs, which involves specific scene descriptions and corresponding rules to generate targeted text descriptions. During the TSCLIP fine-tuning process, we implement adaptive dynamic weight ensembling (ADWE) to seamlessly incorporate outcomes from each training iteration with the zero-shot CLIP model. This approach ensures that the model retains its ability to generalize while acquiring new knowledge about traffic signs. To the best knowledge of authors, TSCLIP is the first contrastive language-image model used for the worldwide cross-regional traffic sign recognition task. The project website is available at: https://github.com/guoyangzhao/TSCLIP.

cs.CV↗

arXiv2024

Algorithm-hardware co-design for Energy-Efficient A/D conversion in ReRAM-based accelerators

Deep neural networks are widely deployed in many fields. Due to the in-situ computation (known as processing in memory) capacity of the Resistive Random Access Memory (ReRAM) crossbar, ReRAM-based accelerator shows potential in accelerating DNN with low power and high performance. However, despite power advantage, such kind of accelerators suffer from the high power consumption of peripheral circuits, especially Analog-to-Digital Converter (ADC), which account for over 60 percent of total power consumption. This problem hinders the ReRAM-based accelerator to achieve higher efficiency. Some redundant Analog-to-Digital conversion operations have no contribution to maintaining inference accuracy, and such operations can be eliminated by modifying the ADC searching logic. Based on such observations, we propose an algorithm-hardware co-design method and explore the co-design approach in both hardware design and quantization algorithms. Firstly, we focus on the distribution output along the crossbar's bit-lines and identify the fine-grained redundant ADC sampling bits. % of weight and To further compress ADC bits, we propose a hardware-friendly quantization method and coding scheme, in which different quantization strategy was applied to the partial results in different intervals. To support the two features above, we propose a lightweight architectural design based on SAR-ADC\@. It's worth mentioning that our method is not only more energy efficient but also retains the flexibility of the algorithm. Experiments demonstrate that our method can reduce about $1.6 \sim 2.3 \times$ ADC power reduction.

cs.AR↗

arXiv2021

PTQ-SL: Exploring the Sub-layerwise Post-training Quantization

Network quantization is a powerful technique to compress convolutional neural networks. The quantization granularity determines how to share the scaling factors in weights, which affects the performance of network quantization. Most existing approaches share the scaling factors layerwisely or channelwisely for quantization of convolutional layers. Channelwise quantization and layerwise quantization have been widely used in various applications. However, other quantization granularities are rarely explored. In this paper, we will explore the sub-layerwise granularity that shares the scaling factor across multiple input and output channels. We propose an efficient post-training quantization method in sub-layerwise granularity (PTQ-SL). Then we systematically experiment on various granularities and observe that the prediction accuracy of the quantized neural network has a strong correlation with the granularity. Moreover, we find that adjusting the position of the channels can improve the performance of sub-layerwise quantization. Therefore, we propose a method to reorder the channels for sub-layerwise quantization. The experiments demonstrate that the sub-layerwise quantization with appropriate channel reordering can outperform the channelwise quantization.

cs.CV↗

arXiv2021

Label-Based Diversity Measure Among Hidden Units of Deep Neural Networks: A Regularization Method

Although the deep structure guarantees the powerful expressivity of deep networks (DNNs), it also triggers serious overfitting problem. To improve the generalization capacity of DNNs, many strategies were developed to improve the diversity among hidden units. However, most of these strategies are empirical and heuristic in absence of either a theoretical derivation of the diversity measure or a clear connection from the diversity to the generalization capacity. In this paper, from an information theoretic perspective, we introduce a new definition of redundancy to describe the diversity of hidden units under supervised learning settings by formalizing the effect of hidden layers on the generalization capacity as the mutual information. We prove an opposite relationship existing between the defined redundancy and the generalization capacity, i.e., the decrease of redundancy generally improving the generalization capacity. The experiments show that the DNNs using the redundancy as the regularizer can effectively reduce the overfitting and decrease the generalization error, which well supports above points.

cs.LG↗

arXiv2021

End-to-End Human Object Interaction Detection with HOI Transformer

We propose HOI Transformer to tackle human object interaction (HOI) detection in an end-to-end manner. Current approaches either decouple HOI task into separated stages of object detection and interaction classification or introduce surrogate interaction problem. In contrast, our method, named HOI Transformer, streamlines the HOI pipeline by eliminating the need for many hand-designed components. HOI Transformer reasons about the relations of objects and humans from global image context and directly predicts HOI instances in parallel. A quintuple matching loss is introduced to force HOI predictions in a unified way. Our method is conceptually much simpler and demonstrates improved accuracy. Without bells and whistles, HOI Transformer achieves $26.61\% $ $ AP $ on HICO-DET and $52.9\%$ $AP_{role}$ on V-COCO, surpassing previous methods with the advantage of being much simpler. We hope our approach will serve as a simple and effective alternative for HOI tasks. Code is available at https://github.com/bbepoch/HoiTransformer .

cs.CV↗

arXiv2020

Quantum observation scheme universally identifying causalities from correlations

It has long been recognized as a difficult problem to determine whether the observed statistical correlation between two classical variables arise from causality or from common causes. Recent research has shown that in quantum theoretical framework, the mechanisms of entanglement and quantum coherence provide an advantage in tackling this problem. In some particular cases, quantum common causes and quantum causality can be effectively distinguished using observations only. However, these solutions do not apply to all cases. There still exist enormous cases in which quantum common causes and quantum causality can not be distinguished. In this paper, along the line of considering unitary transformation as causality in the quantum world, we formally show quantum common causes and quantum causality are universally separable. Based on the analysis, we further provide a general method to discriminate the two.

quant-ph↗

arXiv2020

The optimal way of folding a bond notebook page into a bookmark

The article explores a simple yet interesting geometric problem: how to fold a notebook page so it is the most visible when the notebook is closed (i.e., as a bookmark). We start with a square page, then proceed to the rectangular page. When an additional constraint is added to limit the vertical extent of the rectangular page, the problem shows the first-order phase transition: the optimal way of folding changes drastically when the aspect ratio of the page exceeds the critical value of about $1.20711$.

math.HO↗

arXiv2020

A Geometric Interpretation of the Quadratic Formula

This article provides a simple geometric interpretation of the quadratic formula. The geometry helps to demystify the formula's complex appearance and casts it into a much simpler existence, thus potentially benefits early algebra students.

math.HO↗