Direction, evidence grade, and study type are AI-generated labels (gpt-5-mini), not human-verified. Syntheses are LLM-written. "Tensions" are machine-detected candidates, not confirmed contradictions. A research-acceleration tool, not peer review.
How this is built →
1Distinct papers
22Unique collaborators
1/1Semantic Scholar citation coverage
Publication span: 2026. Corpus fetch span: 2026.
Identity provenance
Provider IDs
- Semantic Scholar:
2405815296
ORCID evidence
No valid ORCID is stored.
Observed aliases (1)
- Tristan Rice (semantic scholar, provider refresh)
Topics and outcomes in this view
Assessment themes
- Adoption: 1 paper
- Innovation: 1 paper
- Org Design: 1 paper
- Productivity: 1 paper
Claim outcomes
- Organizational Efficiency: 1 paper
- Adoption Rate: 1 paper
- Task Completion Time: 1 paper
- Firm Productivity: 1 paper
- Other: 1 paper
- Output Quality: 1 paper
Papers in the Semantic Scholar view
Latest stored Semantic Scholar author observations only. Citation counts below are from the same provider and are not combined with other services.
Scroll the table horizontally to see every column.
| Paper | Author evidence | Date | Provider citations |
|---|---|---|---|
| A fault-tolerant training design (FT-HSDP) halves wasted GPU time at extreme scale: by isolating failures to data-parallel replicas and supporting asynchronous rejoin, it cuts recovery stalls from ~10 minutes to ~3 and raises effective utilization from 44% to 80%, substantially lowering cost per trained model.arxiv | Tristan Rice provider id |
2026-01-30 | 7 |
Citation observation summary
Semantic Scholar supplied counts for 1 of 1 papers in this view; 0 are missing. The observed paper counts sum to 7 cumulative citations. This is a coverage summary, not an author score or h-index.