PulseAugur / Brief
EN
LIVE 17:05:34

Brief

last 24h
[2/2] 224 sources

Multi-source AI news clustered, deduplicated, and scored 0–100 across authority, cluster strength, headline signal, and time decay.

  1. RESEARCH · arXiv cs.CV English(EN) · · [2 sources]

    Wavelength-Multiplexed 2D Beam Steering via a Passive Diffractive Network

    Researchers have developed a novel passive diffractive optical network capable of wavelength-multiplexed 2D beam steering. This system utilizes deep learning to design cascaded diffractive layers that map distinct wavelengths to specific output angles, enabling arbitrary 2D beam steering without mechanical or electronic components. The framework has been numerically demonstrated to steer beams across 625 wavelength channels and experimentally validated in both terahertz and visible spectrum regimes. AI

    IMPACT This research could lead to more compact and efficient optical systems for communications, imaging, and sensing.

  2. TOOL · arXiv cs.LG English(EN) ·

    Low-rank surrogate modeling and stochastic zero-order optimization for training of neural networks with black-box layers

    Researchers have developed a new framework for training hybrid neural networks that combine digital components with physical, black-box layers. This approach uses stochastic zero-order optimization and a dynamic low-rank surrogate model to enable gradient propagation through non-differentiable physical devices. The method has demonstrated effectiveness across computer vision, audio classification, and language modeling tasks, achieving accuracy comparable to purely digital baselines. AI

    Low-rank surrogate modeling and stochastic zero-order optimization for training of neural networks with black-box layers

    IMPACT Offers a practical pathway for integrating non-differentiable physical components into scalable, end-to-end trainable AI systems.