TAIJI proposes an MCP-based architecture that routes each data modality to specialized small models through semantic operators, with preliminary results showing lower latency and better accuracy than a single GPT-4.1 on three queries.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DB 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
TAIJI: MCP-based Multi-Modal Data Analytics on Data Lakes
TAIJI proposes an MCP-based architecture that routes each data modality to specialized small models through semantic operators, with preliminary results showing lower latency and better accuracy than a single GPT-4.1 on three queries.