arXiv subjects

Zhehao Yu

Publications and source records attributed to Zhehao Yu.

8 recordsLinked to original sources

arXiv2026

Accelerating Visual On-Policy Distillation with Batched Speculative Jacobi Rollouts

Visual on-policy distillation (OPD) improves the training of compact visual autoregressive models by learning from trajectories generated by the current student. However, these online rollouts are still produced token by token with autoregressive decoding, which adds substantial cost to every on-policy training step. Speculative Jacobi Decoding (SJD) provides an alternative because it can process multiple tokens in parallel without an auxiliary draft model, but the original method is designed for single-sequence inference. We introduce HB-SJD, a batched SJD rollout backend for visual OPD. HB-SJD allows each image to advance independently according to its own decoding progress, while images at different sequence positions are still verified in batched model forwards. As images finish, HB-SJD switches between Full and Compact execution to reduce the cost of later rollout rounds. HB-SJD only replaces the student rollout backend and leaves the teacher, distillation objective, and optimization procedure unchanged. Experiments with LlamaGen show that HB-SJD substantially reduces rollout and end-to-end training time while preserving the generation quality of the distilled student.

cs.LG↗

arXiv2026

UI-MOPD: Multi-Platform On-Policy Distillation for Unified GUI Agents

Recent advances in multimodal foundation models and agent systems have driven GUI agents from single-platform task execution toward cross-platform interaction. However, unified multi-platform GUI learning remains challenging: high-quality cross-platform trajectories remain scarce, while platforms share transferable capabilities but differ in action semantics and interaction conventions. Naively mixing supervision or merging specialized models can blur native behaviors and produce imbalanced performance. To address these challenges, we construct Uni-GUI, a high-quality dataset containing nearly 10K executable cross-platform interaction trajectories collected through a unified desktop-mobile harness. Building on Uni-GUI, we propose UI-MOPD, the first framework to introduce multi-teacher on-policy distillation (MOPD) into unified multi-platform GUI agent training. UI-MOPD trains a shared student on its own rollouts and dynamically routes each rollout to the corresponding platform-specialized teacher. At student-visited states, teacher guidance serves as a platform-conditioned behavioral anchor, enabling the integration of complementary desktop and mobile expertise without averaging their distinct interaction conventions. On OSWorld and MobileWorld, UI-MOPD achieves task success rates of 38.2% and 12.0%, respectively, outperforming parameter-matched integration strategies while preserving general GUI grounding. These results demonstrate that multi-teacher on-policy distillation provides an effective approach to building unified cross-platform GUI agents. Project page: https://elispectre.github.io/UI-MOPD/.

cs.CL↗

arXiv2026

Xiaomi-GUI-0 Technical Report

Graphical user interface (GUI) agents build on vision-language models to complete user tasks end-to-end in real applications through interface actions such as tapping, swiping, text entry, and navigation. However, existing GUI agents are trained and evaluated largely on offline trajectories, simulated environments, and standardized benchmarks. These differ substantially from real applications in interface layout, interaction logic, and abnormal-state distribution, and cannot faithfully characterize execution stability in real-world use, where account states, permission dialogs, payment authentication, and risk control continually reshape the state distribution and open a persistent gap between benchmark scores and real usability. To close this gap, we propose Xiaomi-GUI-0, a native multimodal GUI agent for real mobile environments, trained and evaluated within a real-device closed loop. At its core is a real-device-dominant hybrid infrastructure, where physical devices are the primary execution environment and sandboxes provide auxiliary support, so that data collection, training, rollout, and evaluation share an execution distribution close to real deployment. We construct multi-source training data spanning high-frequency head tasks, high-generalization data for long-tail intents, and capability-enhancement data for reflection and memory, and introduce an error-driven data flywheel that turns failure trajectories into corrected actions, reflective explanations, and recovery demonstrations. The model is trained through a progressive three-stage pipeline of supervised fine-tuning, step-level reinforcement learning, and agentic reinforcement learning. Evaluated on public benchmarks and our in-house RealMobile, Xiaomi-GUI-0 achieves 72.0% success on RealMobile and 78.9% on AndroidWorld, while substantially improving execution stability and abnormal-state recognition in real-world tasks.

cs.AI↗

arXiv2026

A Scalable All-to-All Reconfigurable Ising Solver Using Pulsed Time-Division Multiplexing

Physics-based computing platforms, such as those based on the Ising model, are an important pillar of future hardware systems built for the artificial intelligence (AI) era. Such platforms show promise for solving nondeterministic polynomial (NP) time problems that are difficult for traditional processing units to solve efficiently as problem size grows. Here, we present a scalable optoelectronic Ising machine architecture, demonstrated with 64 all-to-all connected spins using pulsed time-division multiplexing. The 65 nm CMOS Ising chip integrates the coupling and nonlinear mechanisms in an active area of 3.1 mm2, eliminating the need for benchtop equipment within the loop. The feedback loop of the Ising machine is closed using a compact high-bandwidth, low-loss optical fiber, seamlessly combining optical scalability with the ultradense reconfigurability of integrated electronics. The chip operates at 1 GHz with 4-bit coupling weights and is benchmarked with NP-complete Boolean satisfiability problems consisting of three literals (3-SAT) and clause-to-variable ratios of 32/32, 40/24, and 48/16. Nanosecond annealing times represent at least a three order-of-magnitude improvement over previously reported all-to-all connected works. Time and energy to solutions for 100% 3-SAT clause accuracy are as low as 7.4 us and 2.9 uJ, respectively, achieving more than an order-of-magnitude decrease in time and energy to solution compared to the state of the art. All-to-all connection is demonstrated using MaxCut problems with 100% graph densities. The chip's ability to effectively solve 2-, 3-SAT, and MaxCut problems highlights its reconfigurability and versatility. Furthermore, combining mature CMOS integration with scalable photonic links allows for significant reduction in computation time and energy, addressing the pressing requirements of AI and future hyperscale datacenters.

physics.optics↗

arXiv2026

S2FT: Parameter-Efficient Fine-Tuning in Sparse Spectrum Domain

Parameter Efficient Fine-Tuning (PEFT) is a key technique for adapting a large pretrained model to downstream tasks by fine-tuning only a small number of parameters. Recent methods based on Fourier transforms have further reduced the fine-tuned parameters scale by only fine-tuning a few spectral coefficients. Its basic assumption is that the weight change \delta W is a spatial-domain matrix with a sparse spectrum. However, in this paper, we observe that the spectrum of weight change is not sparse, but instead distributed like power-uniform. This fact implies that fine-tuning only a few spectral coefficients is insufficient to accurately model the weight change with uniform spectrum. To address this issue, we propose to seek an invertible transformation that can transform a latent spatial-domain matrix with sparse spectrum to the weight change, and then perform PEFT on such sparse spectrum domain with few spectral coefficients, called S2FT. To seek such transformation, we first pre-estimate a coarse weight change as a prior. Then, inspired by that sparse spectrum often correspond to locally smooth spatial structures, we regard this transformation as a row and column rearrangement operation on the pre-estimated weight change that smooth spatial structures while keep the structure information of neurons. Finally, we propose to solve the rearrangement search problem in a simple nearest neighbor search manner, thereby obtaining the invertible transformation. Extensive results show our S2FT achieves superior performance by only using 0.08% training parameters.

cs.CV↗

arXiv2026

SJD-PV: Speculative Jacobi Decoding with Phrase Verification for Autoregressive Image Generation

Autoregressive (AR) image models have recently demonstrated remarkable generative capability, but their sequential nature results in significant inference latency. Existing training-free acceleration methods typically verify tokens independently, overlooking the strong co-occurrence patterns between adjacent visual tokens. This independence assumption often leads to contextual inconsistency and limits decoding efficiency. In this work, we introduce a novel training-free acceleration framework that performs phrase-level speculative verification, enabling the model to jointly validate multiple correlated tokens within each decoding window. To construct such phrase units, we analyze token co-occurrence statistics from the training corpus and group frequently co-occurring tokens into semantically coherent visual phrases. During inference, the proposed phrase-level verification evaluates aggregated likelihood ratios over each phrase, allowing simultaneous acceptance of multiple tokens while preserving generation quality. Extensive experiments on autoregressive text-to-image generation show that our method significantly reduces the number of function evaluations (NFE) and achieves up to 30% faster decoding without compromising visual fidelity. Our findings reveal that modeling short-range token co-occurrence provides an effective and general principle for accelerating autoregressive inference.

cs.CV↗

arXiv2026

Single-Chip 1.024 Tb/s Optical Receiver for High-Speed Optical links

Integrated optical transceivers, utilizing wavelength-division-multiplexing, offer a path forward for implementation of compact, high-bandwidth and energy-efficient interconnects for future data centers. Here we report the demonstration of a monolithically integrated optical receiver in 45nm CMOS, where efficient multi-layer optical demultiplexing with capacitive tuning, energy efficient electronics and wideband inverse designed grating couplers enable implementation of a 32-channel receiver chip based on wavelength-division multiplexing. The chip operates at an aggregate data-rate of 1.024 Tb/s with all channels operating simultaneously at a data-rate of 32 Gb/s/channel achieving a record energy efficiency of 71 fJ/b, including the power consumption of both the electronic circuitry and the tuning and control of photonic devices, and a record bandwidth density of 4 Tb/s/mm2. The receiver achieves a bit-error-rate below 1E-12 without requiring equalization, error correction or digital processing. Inverse-designed broadband grating couplers provide efficient, low-loss optical coupling into the chip. An on-chip demultiplexer, composed of Mach-Zehnder interferometers (MZIs) and ring resonators, offers a large channel-to-channel isolation sufficient for error-free operation. Capacitive phase shifters embedded within the ring resonators of the demultiplexer are used for wavelength alignment at a zero static power consumption. MZIs and ring-resonators are periodically selected and autonomously locked to the wavelength of the corresponding optical carrier. The implemented monolithic receiver offers a scalable, energy-efficient and reliable solution for the beyond Tb/s optical interconnects.

physics.optics↗

arXiv2021

mmWave Spatial-Temporal Single Harmonic Switching Transmitter Arrays for High back-off Beamforming Efficiency

This paper presents a spatial-temporal single harmonic switching (STHS) transmitter array architecture with enhanced efficiency in the power back-off (PBO) region. STHS is an electromagnetic and circuit co-designed and jointly optimized transmitter array that realizes beamforming and back-off power generation at the same time. The temporal dimension is originally added in STHS to achieve back-off efficiency enhancement, which can be combined with conventional power back-off enhancement methods such as Doherty amplifiers and envelope tracking. The design is validated through a simulation of a two-stage power amplifier in 65-nm CMOS at 77 GHz, which achieves a peak drain efficiency (DE) of 24.2%, a 22% DE at 3-dB PBO, 16% DE at 6-dB PBO, and 10.2% at 9-dB PBO. The efficiency exhibits a 57% improvement at 3-dB PBO, 100% improvement at 6-dB PBO, and 190% improvement at 9-dB PBO compared with class A/B amplifier.

eess.SP↗