[KongchangAI]
ConceptPersistent Retrieval Sub-agent

Persistent Retrieval Sub-agent

A retrieval architecture for multi-turn dialogue in which a sub-agent maintains retrieval history and state across turns and returns raw retrieved evidence directly to the main agent rather than generating a synthesized answer. The main agent handles evidence evaluation and final generation, decoupling retrieval from reasoning.

Timeline (last 90 days)

Nov 6

A study compared three conversational agent retrieval designs: skill-based main-agent retrieval, non-persistent retrieval sub-agent, and persistent retrieval sub-agent.

Unverified50%
Nov 6

The core of the persistent retrieval sub-agent design is maintaining retrieval history across turns and returning retrieved evidence directly to the main agent, rather than having the sub-agent generate a synthesized response.

Unverified50%
Nov 6

The persistent retrieval design allows the main agent to request corrected retrieval results from the same sub-agent.

Unverified50%
Nov 6

Compared with the non-persistent retrieval sub-agent, the persistent design reduced response time per question by 15.6 seconds.

Unverified50%
Nov 6

Compared with the non-persistent retrieval sub-agent, the persistent design used 65.1% more tokens and had an estimated API cost 29.1% higher.

Unverified50%
Nov 6

Compared with skill-based main-agent retrieval, the persistent retrieval design used 43.3% fewer logged agent-model tokens, reduced estimated API cost by 45.0%, and cut average response time by 9.1%.

Unverified50%
Nov 6

The study evaluated the three designs on enterprise questions, covering standalone questions, questions in shared conversations, and context-dependent follow-up questions, with evaluation spanning five multi-turn settings.

Unverified50%
Nov 6

The persistent retrieval design outperforms skill-based main-agent retrieval in reducing token usage and API cost, while also improving response speed.

Unverified50%
Nov 6

One limitation of the study is that the experiments focus on enterprise questions, and applicability to other domains (such as open-domain dialogue) remains to be verified.

Unverified50%
Nov 6

The increase in token usage under the persistent design may pose challenges for some applications, requiring further optimization of methods for compressing and representing retrieval history.

Unverified50%

1 more timeline events

All Facts (11)

Unverified

A study compared three conversational agent retrieval designs: skill-based main-agent retrieval, non-persistent retrieval sub-agent, and persistent retrieval sub-agent.

50%
Unverified

The core of the persistent retrieval sub-agent design is maintaining retrieval history across turns and returning retrieved evidence directly to the main agent, rather than having the sub-agent generate a synthesized response.

50%
Unverified

The persistent retrieval design allows the main agent to request corrected retrieval results from the same sub-agent.

50%
Unverified

Compared with the non-persistent retrieval sub-agent, the persistent design reduced response time per question by 15.6 seconds.

50%
Unverified

Compared with the non-persistent retrieval sub-agent, the persistent design used 65.1% more tokens and had an estimated API cost 29.1% higher.

50%
Unverified

Compared with skill-based main-agent retrieval, the persistent retrieval design used 43.3% fewer logged agent-model tokens, reduced estimated API cost by 45.0%, and cut average response time by 9.1%.

50%
Unverified

The study evaluated the three designs on enterprise questions, covering standalone questions, questions in shared conversations, and context-dependent follow-up questions, with evaluation spanning five multi-turn settings.

50%
Unverified

The persistent retrieval design outperforms skill-based main-agent retrieval in reducing token usage and API cost, while also improving response speed.

50%
Unverified

One limitation of the study is that the experiments focus on enterprise questions, and applicability to other domains (such as open-domain dialogue) remains to be verified.

50%
Unverified

The increase in token usage under the persistent design may pose challenges for some applications, requiring further optimization of methods for compressing and representing retrieval history.

50%
Unverified

Future research could explore dynamically adjusting retrieval strategies, combining the advantages of persistent and non-persistent approaches, and extending persistent retrieval to multimodal settings.

50%

Source Articles