arXiv subjects

Jiale Yu

Publications and source records attributed to Jiale Yu.

At least 19 recordsLinked to original sources

arXiv2026

Accuracy-Configurable Floating-Point Multiplier Design for SRAM-Based Compute-in-Memory

Digital Compute-in-Memory (DCiM) reduces data movement and has become a promising solution for energy-efficient edge AI. However, most existing DCiM frameworks still primarily target integer or fixed-point arithmetic, and provide limited support for compiler-integrated and accuracy-configurable floating-point computation. Directly integrating conventional IEEE 754 floating-point units into dense SRAM-based DCiM arrays, however, incurs high area and power overhead. To address this challenge, this work presents an accuracy-configurable floating-point multiplier integrated into the OpenACM framework for SRAM-based DCiM. An exact IEEE~754-compliant multiplier is first implemented as a baseline, and a mantissa-segmentation-based approximate multiplier is then proposed to reduce hardware cost while preserving numerical fidelity. Post-layout results show up to 69% logic area reduction and 72% power savings over exact floating-point designs without delay overhead. Evaluations on image processing tasks and ResNet-18 inference further demonstrate negligible accuracy degradation. These results indicate that compiler-integrated approximate floating-point multiplication is a practical approach for enabling efficient and configurable floating-point support in SRAM-based DCiM systems. The Floating-Point Multiplier is available on https://github.com/ShenShan123/OpenACM

cs.AR↗

arXiv2026

HarmoWAM: Harmonizing Generalizable and Precise Manipulation via Adaptive World Action Models

World Action Models (WAMs) have emerged as a promising paradigm for robot control by modeling physical dynamics. Current WAMs generally follow two paradigms: the "Imagine-then-Execute" approach, which uses video prediction to infer actions via inverse dynamics, and the "Joint Modeling" approach, which jointly models actions and video representations. Based on systematic experiments, we observe a fundamental trade-off between these paradigms: the former explicitly leverages world models for generalizable transit but lacks interaction precision, whereas the latter enables fine-grained, temporally coherent action generation but is constrained by the exploration space of the training distribution. Motivated by these findings, we propose HarmoWAM, an end-to-end WAM that fully leverages a world model to unify predictive and reactive control, enabling both generalizable transit and precise manipulation. Specifically, the world model provides spatio-temporal physical priors that condition two complementary action experts: a predictive expert that leverages latent dynamics for iterative action generation, and a reactive expert that directly infers actions from predicted visual evolution. To enable adaptive coordination, a Process-Adaptive Gating Mechanism is proposed to automatically determine the timing and location of switching between them. This allows the world model to drive the reactive expert to expand the exploration space and the predictive expert to perform precise interactions across different stages of a task. For evaluation, we construct three training-unseen test environments across six real-world robotic tasks, covering variations in background, position, and object semantics. Notably, HarmoWAM achieves strong zero-shot generalization across these scenarios, significantly outperforming prior state-of-the-art VLA models and WAMs by margins of 33% and 29%, respectively.

cs.RO↗

arXiv2026

Look Before Acting: Enhancing Vision Foundation Representations for Vision-Language-Action Models

Vision-Language-Action (VLA) models have recently emerged as a promising paradigm for robotic manipulation, in which reliable action prediction critically depends on accurately interpreting and integrating visual observations conditioned on language instructions. Although recent works have sought to enhance the visual capabilities of VLA models, most approaches treat the LLM backbone as a black box, providing limited insight into how visual information is grounded into action generation. Therefore, we perform a systematic analysis of multiple VLA models across different action-generation paradigms and observe that sensitivity to visual tokens progressively decreases in deeper layers during action generation. Motivated by this observation, we propose \textbf{DeepVision-VLA}, built on a \textbf{Vision-Language Mixture-of-Transformers (VL-MoT)} framework. This framework enables shared attention between the vision foundation model and the VLA backbone, injecting multi-level visual features from the vision expert into deeper layers of the VLA backbone to enhance visual representations for precise and complex manipulation. In addition, we introduce \textbf{Action-Guided Visual Pruning (AGVP)}, which leverages shallow-layer attention to prune irrelevant visual tokens while preserving task-relevant ones, reinforcing critical visual cues for manipulation with minimal computational overhead. DeepVision-VLA outperforms prior state-of-the-art methods by 9.0\% and 7.5\% on simulated and real-world tasks, respectively, providing new insights for the design of visually enhanced VLA models.

cs.CV↗

arXiv2026

LaST$_{0}$: Latent Spatio-Temporal Chain-of-Thought for Robotic Vision-Language-Action Model

Vision-Language-Action (VLA) models have recently shown strong generalization, with some approaches seeking to explicitly generate linguistic reasoning traces or predict future observations prior to execution. However, explicit reasoning typically incurs non-negligible inference latency, which constrains the temporal resolution required for robotic manipulation. Moreover, such reasoning is confined to the linguistic space, imposing a representational bottleneck that struggles to faithfully capture ineffable physical attributes. To mitigate these limitations, we propose LaST$_0$, a framework that enables efficient reasoning before acting through a Latent Spatio-Temporal Chain-of-Thought (CoT), capturing fine-grained physical and robotic dynamics that are often difficult to verbalize. Specifically, we introduce a token-efficient latent CoT space that models future visual dynamics, 3D structural information, and robot proprioceptive states, and further extends these representations across time to enable temporally consistent implicit reasoning trajectories. Furthermore, LaST$_0$ adopts a dual-system architecture implemented via a Mixture-of-Transformers design, where a reasoning expert conducts low-frequency latent inference and an acting expert generates high-frequency actions conditioned on robotics-oriented latent representations. To facilitate coordination, LaST$_0$ is trained with heterogeneous operation frequencies, enabling adaptive switching during deployment. Across 10 real-world tasks spanning tabletop, mobile, and dexterous hand manipulation, LaST$_0$ improves mean success rates by 13%, 14% and 14% over prior SOTA VLA methods, respectively.

cs.RO↗

arXiv2025

Towards Cellular-Scale Interpretability in Pathology Foundation Models for Biomarker Assessment

Molecular biomarker testing in pathology is often costly and tissue-consuming, limiting scalable clinical deployment. Artificial intelligence applied to hematoxylin and eosin (HE)-stained histology could enable rapid biomarker screening, but clinical translation requires models that are both accurate and interpretable. Here we introduce Hireca, a biomarker-focused pathology foundation model pretrained on more than 80,000 whole-slide images spanning 38 organ types from three medical centers, together with CytoMap, an interpretability module that localizes cellular-scale evidence underlying predictions. Across 10 biomarker tasks encompassing morphological, molecular, genetic, and spatial-transcriptomic-proxy readouts, Hireca ranked first in five tasks and outperformed comparable models overall. In evaluation by eight pathologists from two countries, CytoMap was consistently preferred over alternative visualization approaches and revealed error patterns in difficult cases. These results position Hireca and CytoMap as a transparent framework for clinically reviewable biomarker assessment directly from routine HE histology.

cs.CV↗

arXiv2025

SimpleVLA-RL: Scaling VLA Training via Reinforcement Learning

Vision-Language-Action (VLA) models have recently emerged as a powerful paradigm for robotic manipulation. Despite substantial progress enabled by large-scale pretraining and supervised fine-tuning (SFT), these models face two fundamental challenges: (i) the scarcity and high cost of large-scale human-operated robotic trajectories required for SFT scaling, and (ii) limited generalization to tasks involving distribution shift. Recent breakthroughs in Large Reasoning Models (LRMs) demonstrate that reinforcement learning (RL) can dramatically enhance step-by-step reasoning capabilities, raising a natural question: Can RL similarly improve the long-horizon step-by-step action planning of VLA? In this work, we introduce SimpleVLA-RL, an efficient RL framework tailored for VLA models. Building upon veRL, we introduce VLA-specific trajectory sampling, scalable parallelization, multi-environment rendering, and optimized loss computation. When applied to OpenVLA-OFT, SimpleVLA-RL achieves SoTA performance on LIBERO and even outperforms $\pi_0$ on RoboTwin 1.0\&2.0 with the exploration-enhancing strategies we introduce. SimpleVLA-RL not only reduces dependence on large-scale data and enables robust generalization, but also remarkably surpasses SFT in real-world tasks. Moreover, we identify a novel phenomenon ``pushcut'' during RL training, wherein the policy discovers previously unseen patterns beyond those seen in the previous training process. Github: https://github.com/PRIME-RL/SimpleVLA-RL

cs.RO↗

arXiv2025

GROVE: A Generalized Reward for Learning Open-Vocabulary Physical Skill

Learning open-vocabulary physical skills for simulated agents presents a significant challenge in artificial intelligence. Current reinforcement learning approaches face critical limitations: manually designed rewards lack scalability across diverse tasks, while demonstration-based methods struggle to generalize beyond their training distribution. We introduce GROVE, a generalized reward framework that enables open-vocabulary physical skill learning without manual engineering or task-specific demonstrations. Our key insight is that Large Language Models(LLMs) and Vision Language Models(VLMs) provide complementary guidance -- LLMs generate precise physical constraints capturing task requirements, while VLMs evaluate motion semantics and naturalness. Through an iterative design process, VLM-based feedback continuously refines LLM-generated constraints, creating a self-improving reward system. To bridge the domain gap between simulation and natural images, we develop Pose2CLIP, a lightweight mapper that efficiently projects agent poses directly into semantic feature space without computationally expensive rendering. Extensive experiments across diverse embodiments and learning paradigms demonstrate GROVE's effectiveness, achieving 22.2% higher motion naturalness and 25.7% better task completion scores while training 8.4x faster than previous methods. These results establish a new foundation for scalable physical skill acquisition in simulated environments.

cs.CV↗

arXiv2023

Logical Magic State Preparation with Fidelity Beyond the Distillation Threshold on a Superconducting Quantum Processor

Fault-tolerant quantum computing based on surface code has emerged as an attractive candidate for practical large-scale quantum computers to achieve robust noise resistance. To achieve universality, magic states preparation is a commonly approach for introducing non-Clifford gates. Here, we present a hardware-efficient and scalable protocol for arbitrary logical state preparation for the rotated surface code, and further experimentally implement it on the \textit{Zuchongzhi} 2.1 superconducting quantum processor. An average of \hhl{$0.8983 \pm 0.0002$} logical fidelity at different logical states with distance-three is achieved, \hhl{taking into account both state preparation and measurement errors.} In particular, \hhl{the magic states $|A^{\pi/4}\rangle_L$, $|H\rangle_L$, and $|T\rangle_L$ are prepared non-destructively with logical fidelities of $0.8771 \pm 0.0009 $, $0.9090 \pm 0.0009 $, and $0.8890 \pm 0.0010$, respectively, which are higher than the state distillation protocol threshold, 0.859 (for H-type magic state) and 0.827 (for T -type magic state).} Our work provides a viable and efficient avenue for generating high-fidelity raw logical magic states, which is essential for realizing non-Clifford logical gates in the surface code.

quant-ph↗

arXiv2022

Experimental quantum computational chemistry with optimised unitary coupled cluster ansatz

Quantum computational chemistry has emerged as an important application of quantum computing. Hybrid quantum-classical computing methods, such as variational quantum eigensolvers (VQE), have been designed as promising solutions to quantum chemistry problems, yet challenges due to theoretical complexity and experimental imperfections hinder progress in achieving reliable and accurate results. Experimental works for solving electronic structures are consequently still restricted to nonscalable (hardware efficient) or classically simulable (Hartree-Fock) ansatz, or limited to a few qubits with large errors. The experimental realisation of scalable and high-precision quantum chemistry simulation remains elusive. Here, we address the critical challenges {associated with} solving molecular electronic structures using noisy quantum processors. Our protocol presents significant improvements in the circuit depth and running time, key metrics for chemistry simulation. Through systematic hardware enhancements and the integration of error mitigation techniques, we push forward the limit of experimental quantum computational chemistry and successfully scale up the implementation of VQE with an optimised unitary coupled-cluster ansatz to 12 qubits. We produce high-precision results of the ground-state energy for molecules with error suppression by around two orders of magnitude. We achieve chemical accuracy for H$_2$ at all bond distances and LiH at small bond distances in the experiment, even beyond the two recent concurrent works. Our work demonstrates a feasible path towards a scalable solution to electronic structure calculation, validating the key technological features and identifying future challenges for this goal.

quant-ph↗

arXiv2022

Self-supervised Group Meiosis Contrastive Learning for EEG-Based Emotion Recognition

The progress of EEG-based emotion recognition has received widespread attention from the fields of human-machine interactions and cognitive science in recent years. However, how to recognize emotions with limited labels has become a new research and application bottleneck. To address the issue, this paper proposes a Self-supervised Group Meiosis Contrastive learning framework (SGMC) based on the stimuli consistent EEG signals in human being. In the SGMC, a novel genetics-inspired data augmentation method, named Meiosis, is developed. It takes advantage of the alignment of stimuli among the EEG samples in a group for generating augmented groups by pairing, cross exchanging, and separating. And the model adopts a group projector to extract group-level feature representations from group EEG samples triggered by the same emotion video stimuli. Then contrastive learning is employed to maximize the similarity of group-level representations of augmented groups with the same stimuli. The SGMC achieves the state-of-the-art emotion recognition results on the publicly available DEAP dataset with an accuracy of 94.72% and 95.68% in valence and arousal dimensions, and also reaches competitive performance on the public SEED dataset with an accuracy of 94.04%. It is worthy of noting that the SGMC shows significant performance even when using limited labels. Moreover, the results of feature visualization suggest that the model might have learned video-level emotion-related feature representations to improve emotion recognition. And the effects of group size are further evaluated in the hyper parametric analysis. Finally, a control experiment and ablation study are carried out to examine the rationality of architecture. The code is provided publicly online.

eess.SP↗

arXiv2022

Realization of fast all-microwave CZ gates with a tunable coupler

The development of high-fidelity two-qubit quantum gates is essential for digital quantum computing. Here, we propose and realize an all-microwave parametric Controlled-Z (CZ) gates by coupling strength modulation in a superconducting Transmon qubit system with tunable couplers. After optimizing the design of the tunable coupler together with the control pulse numerically, we experimentally realized a 100 ns CZ gate with high fidelity of 99.38%$ \pm$0.34% and the control error being 0.1%. We note that our CZ gates are not affected by pulse distortion and do not need pulse correction, {providing a solution for the real-time pulse generation in a dynamic quantum feedback circuit}. With the expectation of utilizing our all-microwave control scheme to reduce the number of control lines through frequency multiplexing in the future, our scheme draws a blueprint for the high-integrable quantum hardware design.

quant-ph↗

arXiv2022

Quantum Neuronal Sensing of Quantum Many-Body States on a 61-Qubit Programmable Superconducting Processor

Classifying many-body quantum states with distinct properties and phases of matter is one of the most fundamental tasks in quantum many-body physics. However, due to the exponential complexity that emerges from the enormous numbers of interacting particles, classifying large-scale quantum states has been extremely challenging for classical approaches. Here, we propose a new approach called quantum neuronal sensing. Utilizing a 61 qubit superconducting quantum processor, we show that our scheme can efficiently classify two different types of many-body phenomena: namely the ergodic and localized phases of matter. Our quantum neuronal sensing process allows us to extract the necessary information coming from the statistical characteristics of the eigenspectrum to distinguish these phases of matter by measuring only one qubit. Our work demonstrates the feasibility and scalability of quantum neuronal sensing for near-term quantum processors and opens new avenues for exploring quantum many-body phenomena in larger-scale systems.

quant-ph↗

arXiv2021

Realization of an Error-Correcting Surface Code with Superconducting Qubits

Quantum error correction is a critical technique for transitioning from noisy intermediate-scale quantum (NISQ) devices to fully fledged quantum computers. The surface code, which has a high threshold error rate, is the leading quantum error correction code for two-dimensional grid architecture. So far, the repeated error correction capability of the surface code has not been realized experimentally. Here, we experimentally implement an error-correcting surface code, the distance-3 surface code which consists of 17 qubits, on the \textit{Zuchongzhi} 2.1 superconducting quantum processor. By executing several consecutive error correction cycles, the logical error can be significantly reduced after applying corrections, achieving the repeated error correction of surface code for the first time. This experiment represents a fully functional instance of an error-correcting surface code, providing a key step on the path towards scalable fault-tolerant quantum computing.

quant-ph↗

arXiv2021

Quantum Computational Advantage via 60-Qubit 24-Cycle Random Circuit Sampling

To ensure a long-term quantum computational advantage, the quantum hardware should be upgraded to withstand the competition of continuously improved classical algorithms and hardwares. Here, we demonstrate a superconducting quantum computing systems \textit{Zuchongzhi} 2.1, which has 66 qubits in a two-dimensional array in a tunable coupler architecture. The readout fidelity of \textit{Zuchongzhi} 2.1 is considerably improved to an average of 97.74\%. The more powerful quantum processor enables us to achieve larger-scale random quantum circuit sampling, with a system scale of up to 60 qubits and 24 cycles. The achieved sampling task is about 6 orders of magnitude more difficult than that of Sycamore [Nature \textbf{574}, 505 (2019)] in the classic simulation, and 3 orders of magnitude more difficult than the sampling task on \textit{Zuchongzhi} 2.0 [arXiv:2106.14734 (2021)]. The time consumption of classically simulating random circuit sampling experiment using state-of-the-art classical algorithm and supercomputer is extended to tens of thousands of years (about $4.8\times 10^4$ years), while \textit{Zuchongzhi} 2.1 only takes about 4.2 hours, thereby significantly enhancing the quantum computational advantage.

quant-ph↗

arXiv2021

Strong quantum computational advantage using a superconducting quantum processor

Scaling up to a large number of qubits with high-precision control is essential in the demonstrations of quantum computational advantage to exponentially outpace the classical hardware and algorithmic improvements. Here, we develop a two-dimensional programmable superconducting quantum processor, \textit{Zuchongzhi}, which is composed of 66 functional qubits in a tunable coupling architecture. To characterize the performance of the whole system, we perform random quantum circuits sampling for benchmarking, up to a system size of 56 qubits and 20 cycles. The computational cost of the classical simulation of this task is estimated to be 2-3 orders of magnitude higher than the previous work on 53-qubit Sycamore processor [Nature \textbf{574}, 505 (2019)]. We estimate that the sampling task finished by \textit{Zuchongzhi} in about 1.2 hours will take the most powerful supercomputer at least 8 years. Our work establishes an unambiguous quantum computational advantage that is infeasible for classical computation in a reasonable amount of time. The high-precision and programmable quantum computing platform opens a new door to explore novel many-body phenomena and implement complex quantum algorithms.

quant-ph↗

arXiv2021

Observation of strong and weak thermalization in a superconducting quantum processor

We experimentally study the ergodic dynamics of a 1D array of 12 superconducting qubits with a transverse field, and identify the regimes of strong and weak thermalization with different initial states. We observe convergence of the local observable to its thermal expectation value in the strong-thermalizaion regime. For weak thermalization, the dynamics of local observable exhibits an oscillation around the thermal value, which can only be attained by the time average. We also demonstrate that the entanglement entropy and concurrence can characterize the regimes of strong and weak thermalization. Our work provides an essential step towards a generic understanding of thermalization in quantum systems.

quant-ph↗

arXiv2021

Quantum walks on a programmable two-dimensional 62-qubit superconducting processor

Quantum walks are the quantum mechanical analogue of classical random walks and an extremely powerful tool in quantum simulations, quantum search algorithms, and even for universal quantum computing. In our work, we have designed and fabricated an 8x8 two-dimensional square superconducting qubit array composed of 62 functional qubits. We used this device to demonstrate high fidelity single and two particle quantum walks. Furthermore, with the high programmability of the quantum processor, we implemented a Mach-Zehnder interferometer where the quantum walker coherently traverses in two paths before interfering and exiting. By tuning the disorders on the evolution paths, we observed interference fringes with single and double walkers. Our work is an essential milestone in the field, brings future larger scale quantum applications closer to realization on these noisy intermediate-scale quantum processors.

quant-ph↗

arXiv2021

Observation of thermalization and information scrambling in a superconducting quantum processor

Understanding various phenomena in non-equilibrium dynamics of closed quantum many-body systems, such as quantum thermalization, information scrambling, and nonergodic dynamics, is a crucial for modern physics. Using a ladder-type superconducting quantum processor, we perform analog quantum simulations of both the $XX$ ladder and one-dimensional (1D) $XX$ model. By measuring the dynamics of local observables, entanglement entropy and tripartite mutual information, we signal quantum thermalization and information scrambling in the $XX$ ladder. In contrast, we show that the $XX$ chain, as free fermions on a 1D lattice, fails to thermalize, and local information does not scramble in the integrable channel. Our experiments reveal ergodicity and scrambling in the controllable qubit ladder, and opens the door to further investigations on the thermodynamics and chaos in quantum many-body systems.

quant-ph↗