UCLA researcher at the intersection of edge-AI benchmarking for foundation models and indirect-sensing vehicle tracking
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
0
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—
Not scored: the OpenAlex snapshot records h-index and i10-index as 0 with sparse yearly counts, indicating insufficient publication-volume data for this identity. This axis would measure aggregate works and per-year productivity.
Impact—
Not scored: citation data is unavailable for this identity — the snapshot lists h-index 0 and i10-index 0, which we treat as missing rather than zero. This axis would measure the citation-based influence of publications.
Focus0/100
Scored 0/100, consistent with the snapshot's field distribution: an even 50/50 split between edge-AI benchmarking/embedded deployment and cyber-physical vehicle tracking via encrypted-traffic side channels. Topic counts likewise spread across five clusters — Advanced Neural Network Applications (4), IoT and Edge/Fog Computing (3), and three 2-paper topics spanning surveillance, target tracking, and autonomous vehicles — a deliberately two-front, breadth-oriented portfolio rather than one concentrated line.
Collaboration41/100
Scored 41/100 on modest network breadth: the snapshot records 7 distinct coauthors across 9 works in the last three years. Both catalogued 2026 papers place Srivastava in the last (senior) author position, though not as corresponding author, alongside recurring collaborators Inesh Chakrabarti, Zejun Xiong, Pragya Sharma, Gaofeng Dong, Vamsi Eyunni, and Kang Yang — a small, stable group typical of a senior-supervisor role.
Activity67/100
Scored 67/100, driven by recent output: 9 works within the last three years, including 8 indexed in 2026 — the snapshot's most active year by far, next to single works in 2005, 2015, and 2025. Both catalogued arXiv preprints appeared in 2026, indicating an intensive current publishing phase concentrated in edge-AI benchmarking and cyber-physical sensing.
Reproducibility55/100
Scored 55/100, the mean across the 2 reproducibility-assessed papers in the catalog, both rated medium readiness. GrayTrack's 60% RMSE reduction is demonstrated in simulation only, and HoliBench's cross-platform claims rest on indirect PAL validation with composability evidence from a single device — mid-tier results that could improve with real-world and broader-device validation.
Research summary
Mani Srivastava is a researcher at the University of California, Los Angeles whose recent catalogued work spans two equally weighted fronts: edge-AI benchmarking and embedded deployment of foundation models, and cyber-physical sensing for vehicle tracking via encrypted-traffic side channels — a 50/50 split in the OpenAlex field distribution. In 2026 he appears as last author on two arXiv preprints: HoliBench, a cross-platform benchmarking and deployment toolkit profiling foundation-model accuracy, latency, and energy across 7 platforms for CPS-IoT applications; and GrayTrack, which turns third-party cameras' encrypted traffic into passage events fused by road-constrained particle filtering, cutting tracking RMSE by 60% in simulation. The snapshot credits him with 9 works over the last three years, including 8 indexed in 2026 — its most active year — alongside 7 distinct coauthors, with recurring collaborators including Inesh Chakrabarti, Zejun Xiong, Pragya Sharma, Gaofeng Dong, Vamsi Eyunni, and Kang Yang. Topic clusters center on Advanced Neural Network Applications (4 works), IoT and Edge/Fog Computing (3), plus tracking- and vehicle-themed clusters. Evolve rates reproducibility mid-tier (mean 55 across 2 assessed papers, both medium readiness). Citation-based output and impact metrics are unavailable in this snapshot and are therefore left unscored; the profile is name-resolved from public bibliographic data.
Collaboration pattern
In the catalogued record Srivastava consistently takes the last (senior) author position — on both 2026 arXiv preprints — though without corresponding-author designation. The snapshot counts 7 distinct coauthors over the recent window, and the recurring group is small and stable: Inesh Chakrabarti, Zejun Xiong, Pragya Sharma, Gaofeng Dong, Vamsi Eyunni, and Kang Yang. The pattern suggests hands-on mentorship of a compact team rather than large multi-institution consortia.
Research keywords
Highlights
Cross-platform benchmarking of foundation models for CPS-IoT
Vehicle tracking from third-party sensors' encrypted traffic
Timeline
- 2005Earliest year in the OpenAlex snapshot: 1 indexed publication recorded.
- 20151 indexed publication recorded in the snapshot for this year.
- 20251 indexed publication — the last record before the current active phase.
- 2026Peak year with 8 indexed works; both catalogued preprints (GrayTrack vehicle tracking, HoliBench CPS-IoT benchmarking) appear on arXiv with Srivastava as last author.
Papers in the Evolve catalog
Frequent co-authors
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.
- Bibliometric figures come from an OpenAlex author snapshot only; h-index and i10-index are recorded as 0 with sparse yearly counts, so the output and impact axes are left unscored.
- Identity is name-resolved from public bibliographic data (OpenAlex, ORCID 0000-0002-3782-9192); the sparse per-year record may reflect incomplete author disambiguation rather than full career output.
- Paper-level analysis covers only 2 arXiv preprints from 2026 reviewed by Evolve; peer-reviewed venue records are not represented in this data.
- The reproducibility score (mean 55) rests on only 2 assessed papers and may not generalize to the full body of work.