arXiv subjects

Ryan Chen

Publications and source records attributed to Ryan Chen.

13 recordsLinked to original sources

arXiv2026

Electric field controlled spin transport in a topological insulator interfaced with a ferroelectric antiferromagnet

Topological insulators have been explored extensively for spin-charge interconversion via magnetic interfaces, yet the true response of their spin-charge conversion, particularly in the absence of an external magnetic field, remains to be studied. Here, we report electric-field control of spin-charge conversion in the topological insulator Bi$_2$Te$_3$ with the antiferromagnetic multiferroic BiFeO$_3$, employing a nonlocal spin transport device. A systematic thickness dependence of the spin transport across the interface between Bi$_2$Te$_3$ and BiFeO$_3$ reveals a signature of topological surface-state-dominated spin transport in the bilayer system. The spin-charge conversion remains robust for thicknesses above 10 nm but falls rapidly with reducing thickness and vanishes at 5 nm. This is consistent with the hybridization-induced emergence of a trivial insulating phase, which is supported by the coherency factor estimated from the magnetoconductance of Bi$_2$Te$_3$. These results establish that spin-momentum-locked surface states dominate interfacial spin transport in the decoupled regime. Beyond presenting efficient spin-charge interconversion at an entirely insulating magnetic interface, this work also highlights sputter-deposited Bi$_2$Te$_3$ as a high-quality and scalable platform for integrating quantum materials into devices. The nonlocal spin transport approach presented here provides a simple and direct evidence of spin-charge conversion and opens an efficient and practical pathway toward designing energy-efficient spin-based devices.

cond-mat.mes-hall↗

arXiv2026

All Leaks Count, Some Count More: Interpretable Temporal Contamination Detection and Mitigation in LLM Backtesting

Backtesting LLMs on resolved events assumes models reason only from pre-cutoff knowledge, yet pretrained models inevitably leak post-cutoff knowledge. We introduce a claim-level evaluation framework that decomposes prediction rationales into atomic claims and applies Shapley values to quantify each claim's decision impact, yielding \textbf{Shapley-DCLR} (\textbf{Shapley}-weighted \textbf{D}ecision-\textbf{C}ritical \textbf{L}eakage \textbf{R}ate) -- an interpretable metric measuring what fraction of decision-driving reasoning is contaminated. We further propose \textbf{TimeSPEC} (\textbf{Time}-\textbf{S}upervised \textbf{P}rediction with \textbf{E}xtracted \textbf{C}laims), an inference-time architecture that interleaves temporally-filtered retrieval with claim-level supervision, producing predictions grounded entirely in pre-cutoff evidence. Across three LLMs, the ablation experiments confirm retrieval and supervision are jointly necessary; and a three-task probe further illstrates that the performance cost of temporal enforcement scales with each task's reliance on post-cutoff information.

cs.AI↗

arXiv2026

Evolutionary System Prompt Learning for Reinforcement Learning in LLMs

Building agentic systems that can autonomously self-improve from experience is a longstanding goal of AI. Large language models (LLMs) today primarily self-improve via two mechanisms: self-reflection for context updates, and reinforcement learning (RL) for weight updates. In this work, we propose Evolutionary System Prompt Learning (E-SPL), a method for jointly improving model contexts and model weights. In each RL iteration, E-SPL samples trajectories under multiple system prompts in parallel, then jointly applies RL updates to LLM weights and evolutionary updates to system prompts. System prompts evolve via mutation and crossover, two genetic operators driven by LLM self-reflection; selection is based on relative performance ratings updated across RL iterations. E-SPL encourages a natural division between declarative knowledge encoded in prompts and procedural knowledge encoded in weights, resulting in improved performance across reasoning and agentic tasks. For instance, in an easy-to-hard (AIME $\rightarrow$ BeyondAIME) generalization setting, E-SPL improves RL success rate from 38.8% $\rightarrow$ 45.1% while also outperforming reflective prompt evolution (40.0%). Overall, our results demonstrate that RL and system prompt evolution are deeply synergistic, and combining the two yields consistent gains in sample efficiency and generalization. Code: https://github.com/LunjunZhang/E-SPL

cs.AI↗

arXiv2025

AIA Forecaster: Technical Report

This technical report describes the AIA Forecaster, a Large Language Model (LLM)-based system for judgmental forecasting using unstructured data. The AIA Forecaster approach combines three core elements: agentic search over high-quality news sources, a supervisor agent that reconciles disparate forecasts for the same event, and a set of statistical calibration techniques to counter behavioral biases in large language models. On the ForecastBench benchmark (Karger et al., 2024), the AIA Forecaster achieves performance equal to human superforecasters, surpassing prior LLM baselines. In addition to reporting on ForecastBench, we also introduce a more challenging forecasting benchmark sourced from liquid prediction markets. While the AIA Forecaster underperforms market consensus on this benchmark, an ensemble combining AIA Forecaster with market consensus outperforms consensus alone, demonstrating that our forecaster provides additive information. Our work establishes a new state of the art in AI forecasting and provides practical, transferable recommendations for future research. To the best of our knowledge, this is the first work that verifiably achieves expert-level forecasting at scale.

cs.AI↗

arXiv2025

LAMP: Extracting Local Decision Surfaces From Large Language Models

We introduce LAMP (Local Attribution Mapping Probe), a method that shines light onto a black-box language model's decision surface and studies how reliably a model maps its stated reasons to its reported predictions by approximating a decision surface. LAMP treats the model's own self-reported explanations as a coordinate system and fits a locally linear surrogate that links those weights to the model's output. By doing so, it reveals how much the stated factors steer the model's decisions. We apply LAMP to three tasks: sentiment analysis, controversial-topic detection, and safety-prompt auditing. Across these tasks, LAMP reveals that many language models' locally approximated linear decision landscapes overall agree with human judgments on explanation quality and, on a clinical case-file data set, align with expert assessments. Since LAMP operates without requiring access to model gradients, logits, or internal activations, it serves as a practical and lightweight framework for auditing proprietary language models, and enabling assessment of whether a model appears to behave consistently with the explanations it provides.

cs.LG↗

arXiv2025

High temperature surface state in Kondo insulator U$_3$Bi$_4$Ni$_3$

The resurgence of interest in Kondo insulators has been driven by two major mysteries: the presence of metallic surface states and the observation of quantum oscillations. To further explore these mysteries, it is crucial to investigate another similar system beyond the two existing ones, SmB$_6$ and YbB$_{12}$. Here, we address this by reporting on a Kondo insulator, U$_3$Bi$_4$Ni$_3$. Our transport measurements reveal that a surface state emerges below 250 K and dominates transport properties below 150 K, which is well above the temperature scale of SmB$_6$ and YbB$_{12}$. At low temperatures, the surface conductivity is about one order of magnitude higher than the bulk. The robustness of the surface state indicates that it is inherently protected. The similarities and differences between U$_3$Bi$_4$Ni$_3$ and the other two Kondo insulators will provide valuable insights into the nature of metallic surface states in Kondo insulators and their interplay with strong electron correlations.

cond-mat.str-el↗

arXiv2024

Why does in-context learning fail sometimes? Evaluating in-context learning on open and closed questions

We measure the performance of in-context learning as a function of task novelty and difficulty for open and closed questions. For that purpose, we created a novel benchmark consisting of hard scientific questions, each paired with a context of various relevancy. We show that counter-intuitively, a context that is more aligned with the topic does not always help more than a less relevant context. This effect is especially visible for open questions and questions of high difficulty or novelty. This result reveals a fundamental difference between the treatment of close-form and open-form questions by large-language models and shows a need for a more robust evaluation of in-context learning on the variety of different types of questions. It also poses a new question of how to optimally select a context for large language models, especially in the context of Retrieval Augmented Generation (RAG) systems. Our results suggest that the answer to this question can be highly application-dependent and might be contingent on factors including the format of the question, the perceived difficulty level of the questions, and the novelty or popularity of the information we seek.

cs.CL↗

arXiv2024

Learning the Market: Sentiment-Based Ensemble Trading Agents

We propose and study the integration of sentiment analysis and deep reinforcement learning ensemble algorithms for stock trading by evaluating strategies capable of dynamically altering their active agent given the concurrent market environment. In particular, we design a simple-yet-effective method for extracting financial sentiment and combine this with improvements on existing trading agents, resulting in a strategy that effectively considers both qualitative market factors and quantitative stock data. We show that our approach results in a strategy that is profitable, robust, and risk-minimal - outperforming the traditional ensemble strategy as well as single agent algorithms and market metrics. Our findings suggest that the conventional practice of switching and reevaluating agents in ensemble every fixed-number of months is sub-optimal, and that a dynamic sentiment-based framework greatly unlocks additional performance. Furthermore, as we have designed our algorithm with simplicity and efficiency in mind, we hypothesize that the transition of our method from historical evaluation towards real-time trading with live data to be relatively simple.

q-fin.TR↗

arXiv2023

A Large-Scale Study of Personal Identifiability of Virtual Reality Motion Over Time

In recent years, social virtual reality (VR), sometimes described as the "metaverse," has become widely available. With its potential comes risks, including risks to privacy. To understand these risks, we study the identifiability of participants' motion in VR in a dataset of 232 VR users with eight weekly sessions of about thirty minutes each, totaling 764 hours of social interaction. The sample is unique as we are able to study the effect of user, session, and time independently. We find that the number of sessions recorded greatly increases identifiability, and duration per session increases identifiability as well, but to a lesser degree. We also find that greater delay between training and testing sessions reduces identifiability. Ultimately, understanding the identifiability of VR activities will help designers, security professionals, and consumer advocates make VR safer.

cs.CR↗

arXiv2020

Bootstrap Confidence Intervals Using the Likelihood Ratio Test in Changepoint Detection

This study aims to evaluate the performance of power in the likelihood ratio test for changepoint detection by bootstrap sampling, and proposes a hypothesis test based on bootstrapped confidence interval lengths. Assuming i.i.d normally distributed errors, and using the bootstrap method, the changepoint sampling distribution is estimated. Furthermore, this study describes a method to estimate a data set with no changepoint to form the null sampling distribution. With the null sampling distribution, and the distribution of the estimated changepoint, critical values and power calculations can be made, over the lengths of confidence intervals.

stat.ME↗

arXiv2020

Observing Responses to the COVID-19 Pandemic using Worldwide Network Cameras

COVID-19 has resulted in a worldwide pandemic, leading to "lockdown" policies and social distancing. The pandemic has profoundly changed the world. Traditional methods for observing these historical events are difficult because sending reporters to areas with many infected people can put the reporters' lives in danger. New technologies are needed for safely observing responses to these policies. This paper reports using thousands of network cameras deployed worldwide for the purpose of witnessing activities in response to the policies. The network cameras can continuously provide real-time visual data (image and video) without human efforts. Thus, network cameras can be utilized to observe activities without risking the lives of reporters. This paper describes a project that uses network cameras to observe responses to governments' policies during the COVID-19 pandemic (March to April in 2020). The project discovers over 30,000 network cameras deployed in 110 countries. A set of computer tools are created to collect visual data from network cameras continuously during the pandemic. This paper describes the methods to discover network cameras on the Internet, the methods to collect and manage data, and preliminary results of data analysis. This project can be the foundation for observing the possible "second wave" in fall 2020. The data may be used for post-pandemic analysis by sociologists, public health experts, and meteorologists.

cs.CY↗

arXiv2018

Lower-Order Biases Second Moments of Dirichlet Coefficients in Families of $L$-Functions

Let $\mathcal E: y^2 = x^3 + A(T)x + B(T)$ be a nontrivial one-parameter family of elliptic curves over $\mathbb{Q}(T)$, with $A(T), B(T) \in \mathbb Z(T)$, and consider the $k$\textsuperscript{th} moments $A_{k,\mathcal{E}}(p) := \sum_{t (p)} a_{\mathcal{E}_t}(p)^k$ of the Dirichlet coefficients $a_{\mathcal{E}_t}(p) := p + 1 - |\mathcal{E}_t (\mathbb{F}_p)|$. Rosen and Silverman proved a conjecture of Nagao relating the first moment $A_{1,\mathcal{E}}(p)$ to the rank of the family over $\mathbb{Q}(T)$, and Michel proved that if $j(T)$ is not constant then the second moment is equal to $A_{2,\mathcal{E}}(p) = p^2 + O(p^{3/2})$. Cohomological arguments show that the lower order terms are of sizes $p^{3/2}, p, p^{1/2}$, and $1$. In every case we are able to analyze in closed form, the largest lower order term in the second moment expansion that does not average to zero is on average negative, though numerics suggest this may fail for families of moderate rank. We prove this Bias Conjecture for several large classes of families, including families with rank, complex multiplication, and constant $j(T)$-invariant. We also study the analogous Bias Conjecture for families of Dirichlet characters, holomorphic forms on GL$(2)/\mathbb{Q}$, and their symmetric powers and Rankin-Selberg convolutions. We identify all lower order terms in large classes of families, shedding light on the arithmetic objects controlling these terms. The negative bias in these lower order terms has implications toward the excess rank conjecture and the behavior of zeros near the central point in families of $L$-functions.

math.NT↗

arXiv2016

On Reay's relaxed Tverberg conjecture and generalizations of Conway's thrackle conjecture

Reay's relaxed Tverberg conjecture and Conway's thrackle conjecture are open problems about the geometry of pairwise intersections. Reay asked for the minimum number of points in Euclidean d-space that guarantees any such point set admits a partition into r parts, any k of whose convex hulls intersect. Here we give new and improved lower bounds for this number, which Reay conjectured to be independent of k. We prove a colored version of Reay's conjecture for k sufficiently large, but nevertheless k independent of dimension d. Requiring convex hulls to intersect pairwise severely restricts combinatorics. This is a higher-dimensional analog of Conway's thrackle conjecture or its linear special case. We thus study convex-geometric and higher-dimensional analogs of the thrackle conjecture alongside Reay's problem and conjecture (and prove in two special cases) that the number of convex sets in the plane is bounded by the total number of vertices they involve whenever there exists a transversal set for their pairwise intersections. We thus isolate a geometric property that leads to bounds as in the thrackle conjecture. We also establish tight bounds for the number of facets of higher-dimensional analogs of linear thrackles and conjecture their continuous generalizations.

math.CO↗