arXiv Science⌕ Search

arXiv subjects

Yuhang Lei

Publications and source records attributed to Yuhang Lei.

5 recordsLinked to original sources

Micius-Flash: Lightweight Agentic Intelligence for Engineering

We introduce Micius-Flash, a 27-billion-parameter model for engineering with native multimodal input, deep reasoning and a 256K-token context window. We design the Micius data pipeline as a knowledge-tree-driven data flywheel comprising three stages: knowledge tree construction and expansion, data synthesis, and quality validation with hard-example augmentation. From this data, we train expert models for semiconductor knowledge and reasoning, hardware coding, and semiconductor multimodal understanding, along with agentic experts. Subsequently, we employ Multi-Teacher On-Policy Distillation (MOPD) to supervise training on sequences generated by Micius itself, thereby integrating the capabilities of these expert models into a unified model. On more than 20 datasets with about 15K records, Micius-Flash averages 64.2% on hardware code benchmarks, 15.0 points above a same-size baseline. It scores 62.4% on a semiconductor-specific agent benchmark, exceeding the same-size baseline by 17.2 points. We further develop Micius Harness for industrial applications, connecting the model to specialized tools, including EDA tools and KiCad, as well as execution environments. Case studies in PCB design, chip design workflow automation, and software development demonstrate the model's ability to use tool execution feedback for multi-step planning, code debugging, and iterative verification. We release Micius as an open-source model to advance the development of agents for science and engineering.

cs.SE↗

On-chip Quantum Measurement of Squeezing Generated from a Silicon Nitride Micro-ring Resonator

Integration of quantum optical technique on-chip is crucial for large scale applications of quantum technology, which were proven in a free space environment to be superior to the corresponding classical technology. Squeezed states of light can be used for enhancing the sensitivity of quantum sensors and for fault-tolerant quantum computing. Although chip-based squeezed light generation has advanced significantly, practical impact remains limited because coupling losses between the chip and off-chip detectors destroy delicate quantum correlations, restricting the amount of observed squeezing. Here, we overcome this limitation by implementing the idea of on-chip quantum measurement with the aid of a parametric amplifier and applying it to the squeezed state generated by a silicon nitride (SiN) microring resonator. In our scheme, two matched SiN micro-rings are sequentially constructed. The first ring generates a squeezed state, whereas the second ring acts as a high-gain parametric amplifier (PA) that measures the squeezed state before the light experiences significant off-chip loss. This architecture is inherently loss-tolerant: the amplifier elevates the quantum noise well above the vacuum level, making the measurement insensitive to downstream losses. We directly observe a quantum noise reduction of 4.6 dB from the first ring, despite a chip-to-fiber coupling loss exceeding 5 dB. This work also demonstrates the first monolithic SU(1,1) interferometer with an estimated 5 dB signal-to-noise enhancement compared to traditional linear interferometers, and thus establishes a practical pathway for chip-based quantum sensors.

quant-ph↗

Generation of a CW anti-bunched photon field from a thin-film PPLN waveguide by two-photon interference with a weak coherent state

An anti-bunched photon field is produced from a thin-film ppln waveguide by mixing the on-chip two-photon state with a weak but matched coherent state. This is achieved by taking out the two-photon part of the coherent state via a destructive two-photon interference with the on-chip two-photon state. We achieve a photon rate of 100 kHz with a g2-value of 0.35. This anti-bunched light field will have applications in high-resolution quantum imaging such as long baseline quantum telescopy for enhancing the signal-to-noise ratio.

quant-ph↗

Single-photon phase spectrum recovery from the Hong-Ou-Mandel dip

Characterizing the temporal-spectral profile of single photons is essential for quantum information protocol utilizing temporal mode for encoding. Based on the phase retrieval algorithm, we present a method to reconstruct the phase spectrum difference between two wave packets from their Hong-Ou-Mandel dip, and intensity spectra. Our confirmatory experiment with weak coherent wave packets demonstrated the accuracy of the reconstructed phase spectrum difference to within plus or minus 0.1 rad. This method is generalizable to the measurement of unknown single-photon wave packets with the aid of a reference wave packet, requiring only the collection of one-dimensional data, which simplifies and expedites the process.

quant-ph↗

Recovery of phase constant from two-photon interference pattern by phase retrieval algorithm

For a HOM interferometer with two independent incident pulses, the interference pattern can be affected by adding a dispersion medium on one of the incident directions, but there hasn't been a method to reconstruct the phase constant of the medium from the interference pattern. To solve it, we adapted two phase retrieval algorithms and used them to recover the phase difference function between the two incident fields, from which the phase constant can be derived. Through simulations, we verified the convergence, accuracy, and robustness of the algorithms, indicating that this phase recovery process can be completed well with negligible error. Our research finds a new application direction for the phase recovery algorithm, provides an algorithmic tool for high-order dispersion measurement using two-photon interference, and paves the way for a higher resolution and phase-sensitive quantum tomography.

quant-ph↗