SearcharxivSearch

arXiv subjects

Yanyu Xiong

Publications and source records attributed to Yanyu Xiong.

3 recordsLinked to original sources

$Q$ Factors Exceeding $10^{4}$ in Wavelength-to-Subwavelength-Scale Free-Space Resonators

Free-space-addressable optical resonators that combine long photon lifetimes (high $Q$ factors) with strong spatial localization of optical fields (small mode volumes, $V_m$) enhance light-matter interactions with facile far-field excitation. The Purcell factor governing spontaneous emission enhancement scales as $Q\,V_m^{-1}$. Periodically asymmetric resonators, in which perturbations convert bound modes into radiating modes, offer a route to free-space resonances, with the radiative $Q$ factor tuned by the geometric and optical strength of the asymmetry-inducing perturbations. However, free-space resonators that simultaneously achieve high $Q$ and small $V_m$ have remained rare. This limitation arises in part because existing designs do not tailor geometric and optical asymmetries concurrently, thus limiting access to high-$Q$ regimes. Here, we show that jointly tuning geometric and optical asymmetries unlocks a biaxial radiative landscape with iso-$Q$ contours that connect disparate perturbations with equivalent $Q$ factors. We demonstrate this framework with very-large-scale-integrated single-crystalline Si nanoantenna pixels (VINPix) with out-of-plane perturbations of 35-150 nm amorphous Si, SiN$_x$, and SiO$_2$. We experimentally establish biaxial $Q$ factor control in air and achieve $Q$ factors up to $76,000$ at wavelength-scale mode volumes ($V_m \sim 1.7\,λ_0^3\,n_{\mathrm{eff}}^{-3}$) in simultaneously imaged arrays of $>80$ resonators in water. Furthermore, we computationally demonstrate 50-nm-wide slotted VINPix that reach $Q$ factors of $10^6$ at subwavelength mode volumes ($V_m \sim 0.2\,λ_0^3\,n_{\mathrm{eff}}^{-3}$) with 20 nm SiO$_2$ perturbations, yielding Purcell factors as high as $5 \times 10^5$ in an all-dielectric free-space resonator.

physics.optics

OpenScholar: Synthesizing Scientific Literature with Retrieval-augmented LMs

Scientific progress depends on researchers' ability to synthesize the growing body of literature. Can large language models (LMs) assist scientists in this task? We introduce OpenScholar, a specialized retrieval-augmented LM that answers scientific queries by identifying relevant passages from 45 million open-access papers and synthesizing citation-backed responses. To evaluate OpenScholar, we develop ScholarQABench, the first large-scale multi-domain benchmark for literature search, comprising 2,967 expert-written queries and 208 long-form answers across computer science, physics, neuroscience, and biomedicine. On ScholarQABench, OpenScholar-8B outperforms GPT-4o by 5% and PaperQA2 by 7% in correctness, despite being a smaller, open model. While GPT4o hallucinates citations 78 to 90% of the time, OpenScholar achieves citation accuracy on par with human experts. OpenScholar's datastore, retriever, and self-feedback inference loop also improves off-the-shelf LMs: for instance, OpenScholar-GPT4o improves GPT-4o's correctness by 12%. In human evaluations, experts preferred OpenScholar-8B and OpenScholar-GPT4o responses over expert-written ones 51% and 70% of the time, respectively, compared to GPT4o's 32%. We open-source all of our code, models, datastore, data and a public demo.

cs.CL

SciCode: A Research Coding Benchmark Curated by Scientists

Since language models (LMs) now outperform average humans on many challenging tasks, it has become increasingly difficult to develop challenging, high-quality, and realistic evaluations. We address this issue by examining LMs' capabilities to generate code for solving real scientific research problems. Incorporating input from scientists and AI researchers in 16 diverse natural science sub-fields, including mathematics, physics, chemistry, biology, and materials science, we created a scientist-curated coding benchmark, SciCode. The problems in SciCode naturally factorize into multiple subproblems, each involving knowledge recall, reasoning, and code synthesis. In total, SciCode contains 338 subproblems decomposed from 80 challenging main problems. It offers optional descriptions specifying useful scientific background information and scientist-annotated gold-standard solutions and test cases for evaluation. Claude3.5-Sonnet, the best-performing model among those tested, can solve only 4.6% of the problems in the most realistic setting. We believe that SciCode demonstrates both contemporary LMs' progress towards becoming helpful scientific assistants and sheds light on the development and evaluation of scientific AI in the future.

cs.AI