arXiv subjects

Sophia Sun

Publications and source records attributed to Sophia Sun.

8 recordsLinked to original sources

arXiv2025

Conformal Prediction for Time-series Forecasting with Change Points

Conformal prediction has been explored as a general and efficient way to provide uncertainty quantification for time series. However, current methods struggle to handle time series data with change points - sudden shifts in the underlying data-generating process. In this paper, we propose a novel Conformal Prediction for Time-series with Change points (CPTC) algorithm, addressing this gap by integrating a model to predict the underlying state with online conformal prediction to model uncertainties in non-stationary time series. We prove CPTC's validity and improved adaptivity in the time series setting under minimum assumptions, and demonstrate CPTC's practical effectiveness on 6 synthetic and real-world datasets, showing improved validity and adaptivity compared to state-of-the-art baselines.

cs.LG↗

arXiv2025

Guardian-regularized Safe Offline Reinforcement Learning for Smart Weaning of Mechanical Circulatory Devices

We study the sequential decision-making problem for automated weaning of mechanical circulatory support (MCS) devices in cardiogenic shock patients. MCS devices are percutaneous micro-axial flow pumps that provide left ventricular unloading and forward blood flow, but current weaning strategies vary significantly across care teams and lack data-driven approaches. Offline reinforcement learning (RL) has proven to be successful in sequential decision-making tasks, but our setting presents challenges for training and evaluating traditional offline RL methods: prohibition of online patient interaction, highly uncertain circulatory dynamics due to concurrent treatments, and limited data availability. We developed an end-to-end machine learning framework with two key contributions (1) Clinically-aware OOD-regularized Model-based Policy Optimization (CORMPO), a density-regularized offline RL algorithm for out-of-distribution suppression that also incorporates clinically-informed reward shaping and (2) a Transformer-based probabilistic digital twin that models MCS circulatory dynamics for policy evaluation with rich physiological and clinical metrics. We prove that \textsf{CORMPO} achieves theoretical performance guarantees under mild assumptions. CORMPO attains a higher reward than the offline RL baselines by 28% and higher scores in clinical metrics by 82.6% on real and synthetic datasets. Our approach offers a principled framework for safe offline policy learning in high-stakes medical applications where domain expertise and safety constraints are essential.

cs.LG↗

arXiv2024

Data-Driven Simulator for Mechanical Circulatory Support with Domain Adversarial Neural Process

Mechanical Circulatory Support (MCS) devices, implemented as a probabilistic deep sequence model. Existing mechanical simulators for MCS rely on oversimplifying assumptions and are insensitive to patient-specific behavior, limiting their applicability to real-world treatment scenarios. To address these shortcomings, our model Domain Adversarial Neural Process (DANP) employs a neural process architecture, allowing it to capture the probabilistic relationship between MCS pump levels and aortic pressure measurements with uncertainty. We use domain adversarial training to combine simulation data with real-world observations, resulting in a more realistic and diverse representation of potential outcomes. Empirical results with an improvement of 19% in non-stationary trend prediction establish DANP as an effective tool for clinicians to understand and make informed decisions regarding MCS patient treatment.

cs.LG↗

arXiv2024

Copula Conformal Prediction for Multi-step Time Series Forecasting

Accurate uncertainty measurement is a key step to building robust and reliable machine learning systems. Conformal prediction is a distribution-free uncertainty quantification algorithm popular for its ease of implementation, statistical coverage guarantees, and versatility for underlying forecasters. However, existing conformal prediction algorithms for time series are limited to single-step prediction without considering the temporal dependency. In this paper, we propose a Copula Conformal Prediction algorithm for multivariate, multi-step Time Series forecasting, CopulaCPTS. We prove that CopulaCPTS has finite sample validity guarantee. On several synthetic and real-world multivariate time series datasets, we show that CopulaCPTS produces more calibrated and sharp confidence intervals for multi-step prediction tasks than existing techniques.

cs.LG↗

arXiv2023

Probabilistic Symmetry for Multi-Agent Dynamics

Learning multi-agent dynamics is a core AI problem with broad applications in robotics and autonomous driving. While most existing works focus on deterministic prediction, producing probabilistic forecasts to quantify uncertainty and assess risks is critical for downstream decision-making tasks such as motion planning and collision avoidance. Multi-agent dynamics often contains internal symmetry. By leveraging symmetry, specifically rotation equivariance, we can improve not only the prediction accuracy but also uncertainty calibration. We introduce Energy Score, a proper scoring rule, to evaluate probabilistic predictions. We propose a novel deep dynamics model, Probabilistic Equivariant Continuous COnvolution (PECCO) for probabilistic prediction of multi-agent trajectories. PECCO extends equivariant continuous convolution to model the joint velocity distribution of multiple agents. It uses dynamics integration to propagate the uncertainty from velocity to position. On both synthetic and real-world datasets, PECCO shows significant improvements in accuracy and calibration compared to non-equivariant baselines.

cs.LG↗

arXiv2022

Conformal Methods for Quantifying Uncertainty in Spatiotemporal Data: A Survey

Machine learning methods are increasingly widely used in high-risk settings such as healthcare, transportation, and finance. In these settings, it is important that a model produces calibrated uncertainty to reflect its own confidence and avoid failures. In this paper we survey recent works on uncertainty quantification (UQ) for deep learning, in particular distribution-free Conformal Prediction method for its mathematical properties and wide applicability. We will cover the theoretical guarantees of conformal methods, introduce techniques that improve calibration and efficiency for UQ in the context of spatiotemporal data, and discuss the role of UQ in the context of safe decision making.

cs.AI↗

arXiv2001

Optical properties of self-assembled quantum wires for application in infra-red detection

The theoretical studies of optical properties of Ga$_{1-x}$In$_x$As self-assembled quantum-wires (QWR's) made of short-period superlattices with strain-induced lateral ordering are presented. Valence-band anisotropy, band mixing, and effects due to local strain distribution at the atomistic level are all taken into account. Using realistic material parameters which are experimentally feasible, we perform simulations of the absorption spectra for both inter-subband and inter-band transitions (including the excitonic effect) of this material.It is shown that the self-assembled QWR's have favorable optical properties for application in infra-red detection with normal incidence. The wavelength of detection ranges from 10 $μm$ to 20 $μm$ with the length of QWR period varying from 140Åto 300Å.

cond-mat.mtrl-sci↗

arXiv2000

Systematic study of Ga$_{1-x}$In$_x$As self-assembled quantum wires with different interfacial strain relaxation

A systematic theoretical study of the electronic and optical properties of Ga$_{1-x}$In$_x$As self-assembled quantum-wires (QWR's) made of short-period superlattices (SPS) with strain-induced lateral ordering is presented. The theory is based on the effective bond-orbital model (EBOM) combined with a valence-force field (VFF) model. Valence-band anisotropy, band mixing, and effects due to local strain distribution at the atomistic level are all taken into account. Several structure models with varying degrees of alloy mixing for lateral modulation are considered. A valence force field model is used to find the equilibrium atomic positions in the QWR structure by minimizing the lattice energy. The strain tensor at each atomic (In or Ga) site is then obtained and included in the calculation of electronic states and optical properties. It is found that different local arrangement of atoms leads to very different strain distribution, which in turn alters the optical properties. In particular, we found that in model structures with thick capping layer the electron and hole are confined in the Ga-rich region and the optical anisotropy can be reversed due to the variation of lateral alloying mixing, while for model structures with thin capping layer the electron and hole are confined in the In-rich region, and the optical anisotropy is much less sensitive to the lateral alloy mixing.

cond-mat.mes-hall↗