Liqiang Nie

Harbin Institute of Technology · China

Unclaimed · AI-structured profileOpenAlex

HIT researcher spanning multimodal agentic retrieval, egocentric 4D forecasting, and aerial vision-language navigation

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-16) 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—

No output score yet: this snapshot lacks reliable per-venue publication volume for the record. This axis would measure total works over time, peer-reviewed venue counts, and first/corresponding-author share, separate from the concentrated 2026 arXiv burst.

Impact—

No impact score yet: citation coverage for this OpenAlex record is sparse (h-index 1, i10-index 0), too thin to score. Impact would track accumulated citations, h-index trajectory, and uptake of the three 2026 preprints by other work.

Focus0/100

Focus scores 0 because OpenAlex splits the profile evenly across three field clusters — robotics/embodied AI (aerial vision-language navigation), multimodal AI agents/information retrieval, and computer vision (egocentric perception and human motion forecasting) — each weighted 0.333. The three reviewed 2026 preprints sit one per cluster, so current work is deliberately broad rather than concentrated in a single line.

Collaboration55/100

Collaboration scores 55, mid-range. The snapshot records 15 distinct coauthors across 60 recent-3y works, and twelve collaborators recur on the frequency list, led by Jianqiang Xiao, Xiang Deng, Yuexuan Sun, and Yanjin Wu. On all three Evolve-reviewed 2026 preprints Nie sits in the last author position, consistent with a senior role inside a stable, moderately sized team rather than a large rotating consortium.

Activity100/100

Activity scores a maximum 100. OpenAlex countsByYear shows 1 work in 2025 surging to 59 in 2026, and recent3yWorks totals 60 — virtually the entire recorded output lands in the current year. Topic counts reinforce the pace, led by Multimodal Machine Learning Applications (36), Robot Manipulation and Learning (16), and Advanced Neural Network Applications (14).

Reproducibility55/100

Reproducibility scores 55, the mean readiness Evolve assigned across 3 assessed papers (reproPapers 3, reproMean 55), each labelled medium. Recurring caveats keep the score from rising: AeroBelief reports the best SR/OSR/SPL on UAV-ON but is simulation-only with no variance reported, while the HIGFlow paper's Health-benchmark gains are judged marginal despite strong ablations.

Research summary

Liqiang Nie is a researcher at Harbin Institute of Technology whose profile is name-resolved from OpenAlex bibliographic data. The snapshot shows explosive, current-year-driven output: 60 works in the last three years, 59 recorded in 2026 against just 1 in 2025, giving a maximal activity score of 100. Work spreads evenly across three clusters — robotics/embodied AI with aerial vision-language navigation, multimodal AI agents and information retrieval, and computer vision centred on egocentric perception and human motion forecasting — each weighted 0.333, which yields a focus score of 0. Topic counts are led by Multimodal Machine Learning Applications (36), Robot Manipulation and Learning (16), Advanced Neural Network Applications (14), GANs and image synthesis (13), and reinforcement learning in robotics (11). Evolve has reviewed three 2026 arXiv preprints, all with Nie in the last author position: MM-ContextFold (training-free dual-state context folding, +6.3pp over ReAct at 27.5% shorter context), a 233K-sample continuous 4D egocentric forecasting dataset with cascaded HIGFlow, and AeroBelief (dual-layer VLM belief mapping for UAV object goal navigation). Collaboration is moderate (score 55; 15 distinct coauthors), reproducibility readiness averages 55 across the three papers, and citation coverage remains sparse (h-index 1, i10-index 0).

Collaboration pattern

Collaboration scores 55 with 15 distinct coauthors across 60 recent-3y works. Twelve collaborators recur on the frequency list, led by Jianqiang Xiao, Xiang Deng, Yuexuan Sun, Yanjin Wu, Wenbiao Yan, and Yang Tian. On all three Evolve-reviewed 2026 preprints Nie holds the last author position without corresponding-author status, suggesting a senior role within a compact, stable team.

Research keywords

multimodal agentic retrievalcontext management for LLM agentsegocentric vision4D interaction forecastingflow matchinghuman motion forecastingaerial object goal navigationvision-language modelssemantic-spatial belief mappingUAV autonomysearch agentsreinforcement learning in robotics

Highlights

Training-free context folding lifts multimodal agentic retrieval
MM-ContextFold reports training-free dual-state context folding that improves multimodal agentic retrieval accuracy by 6.3pp over ReAct while producing 27.5% shorter contexts; Evolve judges the broad evidence mostly convincing.
233K-sample benchmark for continuous 4D egocentric forecasting
From Where to How introduces a 233K-sample continuous 4D egocentric interaction forecasting dataset together with cascaded HIGFlow; Evolve highlights the strong formulation and ablations while noting the Health-benchmark gains are marginal.
AeroBelief posts best reported UAV ObjectNav results
Dual-layer semantic-spatial VLM belief mapping achieves the best reported SR/OSR/SPL on UAV-ON for aerial object goal navigation; Evolve cautions the gains are modest, simulation-only, and reported without variance.

Timeline

2025Earliest recorded OpenAlex activity: 1 work appears in the 2025 countsByYear, before the current surge.
2026Recorded output surges to 59 works in 2026 (of 60 over the last three years), driving the maximal activity score of 100.
2026MM-ContextFold preprint on training-free dual-state context folding for multimodal agentic retrieval, with Nie as last author.
2026Releases a 233K-sample continuous 4D egocentric interaction forecasting dataset with cascaded HIGFlow, last-author position.
2026AeroBelief preprint reports the best SR/OSR/SPL on UAV-ON for aerial object goal navigation, last-author position.

Papers in the Evolve catalog

Frequent co-authors

Jianqiang Xiao×1 Xiang Deng×1 Yuexuan Sun×1 Yanjin Wu×1 Wenbiao Yan×1 Yang Tian×1 Fan Liu×1 Jingyuan Zhang×1 Zhenyang Li×1 Yupeng Hu×1 Qiaohui Chu×1 Haoyu Zhang×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.

Bibliometrics come from OpenAlex only; identity is name-resolved, and Liqiang Nie is a common name — without an ORCID this profile may merge or split records.
Citation coverage is sparse in this snapshot (h-index 1, i10-index 0), so the output and impact axes are left unscored.
The paper-level evidence here is limited to three 2026 arXiv preprints reviewed by Evolve, all in last-author, non-corresponding positions; venue and peer-review status are not in the data.
The reproducibility score averages only 3 assessed papers (reproMean 55) and may shift as more of the catalog is reviewed.