Designing LLM chains by adapting techniques from crowdsourcing workflows

M Grunde-McLaughlin, MS Lam, R Krishna… - arxiv preprint arxiv …, 2023 - arxiv.org
LLM chains enable complex tasks by decomposing work into a sequence of subtasks.
Similarly, the more established techniques of crowdsourcing workflows decompose complex …

The Design of an LLM-powered Unstructured Analytics System

E Anderson, J Fritz, A Lee, B Li, M Lindblad… - arxiv preprint arxiv …, 2024 - arxiv.org
LLMs demonstrate an uncanny ability to process unstructured data, and as such, have the
potential to go beyond search and run complex, semantic analyses at scale. We describe …

LLM-Powered Proactive Data Systems

S Zeighami, Y Lin, S Shankar… - arxiv preprint arxiv …, 2025 - arxiv.org
With the power of LLMs, we now have the ability to query data that was previously
impossible to query, including text, images, and video. However, despite this enormous …

PalimpChat: Declarative and Interactive AI analytics

C Liu, G Vitagliano, B Rose, M Prinz… - arxiv preprint arxiv …, 2025 - arxiv.org
Thanks to the advances in generative architectures and large language models, data
scientists can now code pipelines of machine-learning operations to process large …

[PDF][PDF] VectraFlow: Integrating Vectors into Stream Processing

D Lu, S Feng, J Zhou, F Solleza, M Schwarzkopf… - vldb.org
Vectors have quickly become the de facto standard in modern search-oriented applications
due to their ability to represent complex data in a structured and effective manner. While …

[PDF][PDF] Towards Interventions for Suicide Prevention with LLM Assistants in Social Work

J Ranjit, J Cho, M Phung, JR Blosnich… - jr4fs.github.io
Suicide is one of the leading causes of death in the US: one person dies by suicide every 11
minutes and one is hospitalized for a suicide attempt every 54 seconds (cdc, 2024b) …