arXiv subjects

Peter Zhong

Publications and source records attributed to Peter Zhong.

5 recordsLinked to original sources

arXiv2026

What People See (and Miss) About Generative AI Risks: Perceptions of Failures, Risks, and Who Should Address Them

Despite growing concerns about the risks of Generative AI (GenAI), there is limited understanding of public perceptions of these risks and their associated failure modes -- defined as recurring patterns of sociotechnical breakdown across the GenAI lifecycle that contribute to risks of real-world harm. To address this gap, we present a survey instrument, validated with eight subject matter experts and deployed on a sample of 960 U.S.-based participants, to assess awareness and perceptions of GenAI's failure modes, their associated risks, and stakeholder responsibilities to address them. To support realism and content validity, our instrument is structured around scenarios grounded in publicly reported incidents and a taxonomy of GenAI's failure modes. Findings suggest that our instrument is (1) effective for assessing risk awareness and perceptions in a way that is grounded in people's current contexts of use, yet is extensible to new contexts that will inevitably arise; and (2) potentially useful for informing the design of AI literacy tools and interventions. We argue for AI literacy and governance approaches that align with how people encounter and reason about GenAI in everyday life.

cs.HC↗

arXiv2026

Enhancing LLM Medical Coding with Structured External Knowledge

Accurate medical coding requires consulting authoritative resources such as the ICD tabular list and coding guidelines. Existing LLM-based automated methods largely rely on LLMs' internal knowledge, which is prone to hallucination and cannot keep pace with guideline updates. We introduce RAG-Coding, an agentic, training-free method that augments LLMs with structured external knowledge: the tabular list is encoded as a knowledge graph capturing hierarchical and instructional code relationships, and the guidelines are distilled into concise, code-specific summaries rather than retrieved as raw text. To enable our study, we also introduce MDACE-2025, expert re-annotations of the MDACE dataset under the 2025 ICD-10-CM/PCS guidelines, adding code sequencing and justification comments. On MDACE, RAG-Coding outperforms the best LLM-based baseline by 3--13\% in micro-F1 across five LLM backbones, and achieves comparable micro- and macro-F1 to the supervised state-of-the-art, with higher recall ($+$11\%) at the cost of precision ($-$6\%). On MDACE-2025, RAG-Coding outperforms all baselines, demonstrating effective generalisation to updated guidelines. Ablations confirm stepwise gains, highlighting the importance of integrating structured external knowledge for LLM-based medical coding.

cs.CL↗

arXiv2025

Dynamic spectral tailoring of a 10 GHz laser frequency comb for enhanced calibration of astronomical spectrographs

Laser frequency combs (LFCs) are an important component of Doppler radial velocity (RV) spectroscopy that pushes fractional precision to the $10^{-10}$ level, as required to identify and characterize Earth-like exoplanets. However, large intensity variations across the LFC spectrum that arise in nonlinear broadening limit the range of comb lines that can be used for optimal wavelength calibration with sufficient signal-to-noise ratio. Furthermore, temporal spectral-intensity fluctuations of the LFC, that are coupled to flux-dependent detector defects, alter the instrumental point spread function (PSF) and result in spurious RV shifts. To address these issues and improve calibration precision, spectral flattening is crucial for LFCs to maintain a constant photon flux per comb mode. In this work, we demonstrate a dynamic spectral shaping setup using a spatial light modulator (SLM) over the wavelength range of 800nm to 1300nm. The custom shaping compensates for amplitude fluctuations in real time and can also correct for wavelength-dependent spectrograph transmission, achieving a spectral profile that delivers the constant readout necessary for maximizing precision. Importantly, we characterize the out-of-loop properties of the spectral flattener to verify a twofold improvement in spectral stability. This technique, combined with our approach of pumping the waveguide spectral broadener out-of-band at 1550 nm, reduces the required dynamic range. While this spectral region is tailored for the LFC employed at the Habitable-zone Planet Finder (HPF) spectrograph, the method is broadly applicable to any LFC used for astronomical spectrograph calibration.

astro-ph.IM↗

arXiv2024

A Calculus for Unreachable Code

In Racket, the LLVM IR, Rust, and other modern languages, programmers and static analyses can hint, with special annotations, that certain parts of a program are unreachable. Same as other assumptions about undefined behavior; the compiler assumes these hints are correct and transforms the program aggressively. While compile-time transformations due to undefined behavior often perplex compiler writers and developers, we show that the essence of transformations due to unreachable code can be distilled in a surprisingly small set of simple formal rules. Specifically, following the well-established tradition of understanding linguistic phenomena through calculi, we introduce the first calculus for unreachable. Its term-rewriting rules that take advantage of unreachable fall into two groups. The first group allows the compiler to delete any code downstream of unreachable, and any effect-free code upstream of unreachable. The second group consists of rules that eliminate conditional expressions when one of their branches is unreachable. We show the correctness of the rules with a novel logical relation, and we examine how they correspond to transformations due to unreachable in Racket and LLVM.

cs.PL↗

arXiv2021

Understanding in Artificial Intelligence

Current Artificial Intelligence (AI) methods, most based on deep learning, have facilitated progress in several fields, including computer vision and natural language understanding. The progress of these AI methods is measured using benchmarks designed to solve challenging tasks, such as visual question answering. A question remains of how much understanding is leveraged by these methods and how appropriate are the current benchmarks to measure understanding capabilities. To answer these questions, we have analysed existing benchmarks and their understanding capabilities, defined by a set of understanding capabilities, and current research streams. We show how progress has been made in benchmark development to measure understanding capabilities of AI methods and we review as well how current methods develop understanding capabilities.

cs.AI↗