Yi Li

The University of Texas at Dallas

Unclaimed · AI-structured profileOpenAlex

UT Dallas researcher in agentic memory systems and theory-of-mind auditing for LLM multi-agent social simulation.

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

—

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-13) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.

Research profile

Output—

Not yet scored: this snapshot contains no publication-count or productivity data beyond the two Evolve-reviewed papers, so total output cannot be computed. This axis would measure cumulative publication volume and its distribution across venues and years.

Impact—

Not yet scored: no citation counts, h-index, or venue-level influence data are available in this snapshot. This axis would measure how often the two 2026 preprints and any wider body of work are cited and taken up by other researchers.

Focus0/100

Focus scores 0 because this snapshot shows an even 50/50 split between two fields — AI systems (LLM agents and memory architectures) and AI evaluation / multi-agent systems (LLM social simulation) — with one reviewed paper in each. Rather than deep single-track specialization, the current record describes a two-sided agenda: building agentic memory infrastructure while simultaneously auditing how LLM multi-agent simulations behave.

Collaboration44/100

The collaboration score is 44, drawn from 8 distinct coauthors in this snapshot. Recurring partners include Dongming Jiang, Bingzhe Li, Cong Li, Cheng Chen, Thomas Fung, and Alex Rossi. On the two reviewed 2026 papers Yi Li appears in last and middle author positions without corresponding authorship, suggesting a contributing role on these entries rather than a leading-corresponding one — a modest, working-group scale network so far.

Activity—

Not yet scored: the snapshot offers no multi-year publication cadence — both reviewed papers are dated 2026 — so recent-output trends cannot be assessed. This axis would measure publishing rhythm and ongoing research momentum over time.

Reproducibility55/100

Reproducibility averages 55 across the two reviewed papers, each rated medium readiness. Evolve's review of Jev-Mem credits the headline gains — LoCoMo 0.777, +11% relative, 6.6x faster builds — but flags a single benchmark and missing ablations; the theory-of-mind audit is called sound with a single-reader caveat. The mid-range score reflects solid, checkable claims that still lack the robustness checks needed for confident replication.

Research summary

Yi Li is a researcher at The University of Texas at Dallas whose visible record sits at the intersection of LLM-based agent systems and their evaluation. The Evolve catalog holds two 2026 arXiv preprints, and the field distribution splits evenly at 0.5/0.5 between AI systems (LLM agents and memory architectures) and AI evaluation / multi-agent systems (LLM social simulation), producing a focus score of 0 — a two-sided agenda that both builds agent infrastructure and audits it. The systems thread, Jev-Mem, introduces System-One-controlled agentic memory; Evolve's review reports LoCoMo at 0.777 (+11% relative) with 6.6x faster builds while flagging a single benchmark and no ablations. The evaluation thread, Mind or Message?, audits theory of mind in multi-agent social simulation and finds that negotiation agents track messages rather than minds — fluent but projection-driven and frontier-blind. Reproducibility averages 55 across the two papers, both rated medium readiness. The collaboration score of 44 reflects 8 distinct coauthors, including Dongming Jiang, Bingzhe Li, Cong Li, Cheng Chen, Thomas Fung, and Alex Rossi; Yi Li is last author on one paper and middle author on the other, neither corresponding. Output, impact, and activity lack underlying data and are unscored. With a common name and no ORCID, this profile is name-resolved from public bibliographic data.

Collaboration pattern

8 distinct coauthors appear in this snapshot (collaboration score 44); recurring partners include Dongming Jiang, Bingzhe Li, Cong Li, Cheng Chen, Thomas Fung, and Alex Rossi. Across the two reviewed 2026 papers, Yi Li is last author on the theory-of-mind audit and middle author on Jev-Mem, neither as corresponding author — a contributing role on small-team arXiv work. The raw list repeats “Cheng Chen” and includes “Yi Li” itself, so counts may be noisy.

Research keywords

LLM agentsagentic memorytheory of mindmulti-agent systemssocial simulationmemory efficiencyadaptive retrievalmulti-relational knowledge graphsSystem-One/System-Two cognitionmulti-issue negotiationegocentric projectionAI evaluation and benchmarking

Highlights

Negotiation agents track messages, not minds
In a multi-issue negotiation audit of LLM multi-agent social simulation, Evolve's review reports that agents appear fluent but are projection-driven and frontier-blind — they mirror their own message context rather than modeling partners' minds. The audit itself is judged sound, with a single-reader caveat noted.
System-One control for efficient agentic memory
Jev-Mem applies System-One/System-Two cognition to agentic memory built on multi-relational knowledge graphs with adaptive retrieval. Evolve's review reports LoCoMo at 0.777 (+11% relative) with 6.6x faster builds, judging the efficiency design promising while noting a single benchmark and no ablations.
A two-sided 2026 record: building and auditing LLM agents
Both catalog entries date to 2026 and split evenly between AI systems (LLM agents and memory architectures) and AI evaluation / multi-agent systems (LLM social simulation), each rated medium reproducibility readiness — an early-stage, dual-track record with clear room to grow on both sides.

Timeline

2026Posted the arXiv preprint “Jev-Mem: System-One-Controlled Agentic Memory for Efficient AI Agents” (middle author, AI systems / memory architectures track).
2026Posted the arXiv preprint “Mind or Message? Auditing Theory of Mind in Multi-Agent Social Simulation” (last author, AI evaluation / LLM social simulation track).
2026Both 2026 preprints entered the Evolve catalog, splitting the field distribution evenly at 0.5/0.5, with medium reproducibility readiness on each.

Papers in the Evolve catalog

Frequent co-authors

Dongming Jiang×1 Bingzhe Li×1 Cong Li×1 Cheng Chen×1 Thomas Fung×1 Alex Rossi×1 Cheng Chen×1 Yi Li×1

About this profile

This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.

All metrics come from the provided bibliometric snapshot and Evolve-reviewed papers only; no external sources such as ORCID records or citation databases were available.
“Yi Li” is a common name and no ORCID or country field was provided, so this profile is name-resolved from public bibliographic data.
Qualitative claims rest on only two Evolve-reviewed arXiv preprints, both from 2026; this is an early-stage profile.
The frequent-coauthor list contains repeated entries (“Cheng Chen” twice, plus “Yi Li” itself), so collaboration counts may be slightly inflated or noisy.