Current Position
University of Nevada, Reno
Ph.D. in Statistics and Data Science
Advisor: Dr. Deena Schmidt
01
Research
Graph-Valued Inference
Statistical inference on graph-valued data
Developing multi-factor ANOVA procedures for graph-valued random objects within the metric-space inference framework. A central question is when Fréchet mean-based methods are well-posed and when distance-based procedures are more appropriate given the geometry and computational constraints of the graph space.
Generative Models
Graph generative models and LLM-based graph generation
Building regime-aware evaluation pipelines for graph generative models and comparing classical statistical generators with newer deep approaches. Related work studies LLMs as sequential graph generators, where each emitted edge depends on the preceding graph history.
AI Evaluation
Multi-agent modeling and AI evaluation
Studying multi-agent interaction and AI evaluation when outputs are structured rather than scalar. This includes graph-valued extensions of longitudinal modeling and diagnostic frameworks for AI-mediated learning.
02
Recent Activity
Current Work
Active research and applied work
Statistical inference on graph-valued data 2024 — Present
Developing graph-valued ANOVA and related distance-based inference procedures for non-Euclidean random objects.
LLMs as stochastic graph generators 2025 — Present
Studying sequential edge generation, in-context variation, and mechanistic signatures in LLM-generated graphs.
Multi-agent interaction modeling 2025 — Present
Extending cross-lag panel ideas to multi-agent systems in which each agent produces graph-structured output.
Recent Activity
Talks, workshops, and current dissemination
- Networks for Science Workshop, Berkeley Lab 2025
Poster: A Framework for Quantifying Variability in LLM-Generated Graphs. - Summer Institute in Computational Social Science (SICSS-Penn), University of Pennsylvania 2025
Collaboration on multi-agent LLM frameworks and longitudinal influence modeling. - SIAM Conference on Mathematics of Data Science (MDS24) 2024
Co-organized the mini-symposium Graph Learning and Network Analytics: Framework, Information Flow, and Applications; presented Decoding Social Networks: LLM Agents and Information Flow Dynamics. - First Annual Data Science Conference, University of Nevada, Reno 2024
Poster presenter.
03
Selected Work
Projects
- LLM Agents in Social Network Dynamics: A Study on Information Flow and Propagation
- Exploring Relationships Between Network Properties: A Regression Analysis
- DriveOhio, Autonomous Vehicle Shuttle Service Planning
- Insights into Network Science and Epidemic Modeling in Network Data
- Gold and Bitcoin Portfolio Modeling, 2022 (Meritorious Winner)
Publications
- Lamichhane, B., Joseph Palardy, Singh, A.K. The Nuances of Large-Language-Model-Agent Performance in Simple English Auctions. Empirical Economics Letters, 2024.
- Lamichhane, B., Singh, A.K., Devkota, S., Dhakal, U., Singh, S., and Dhakal, C. Understanding the Influence of Genre-Specific Music Using Network Analysis and Machine Learning Algorithms. MDPI Big Data and Cognitive Computing, 2023.
- Singh, A.K., Devkota, S., Lamichhane, B., Dhakal, U., and Dhakal, C. Do Large Language Models Show Human-like Biases? Exploring the Confidence-Competence Gap in AI. MDPI Information, 2024.
04
Background
Education
- University of Nevada, Reno
Ph.D. in Statistics and Data Science Aug 2022 — Present - Youngstown State University
B.S. in Mathematics, Minor in Data Analytics Aug 2018 — May 2022
Teaching
- STAT 152 — Introduction to Statistics
- MATH 126 / 126E / 126EE — Pre-Calculus
- MATH 182 — Calculus II (recitation)
Links