{"id":1046808,"date":"2024-06-12T22:48:19","date_gmt":"2024-06-13T05:48:19","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?post_type=msr-project&p=1046808"},"modified":"2024-06-12T22:48:21","modified_gmt":"2024-06-13T05:48:21","slug":"uniprompt-a-structured-approach-to-prompt-optimization","status":"publish","type":"msr-project","link":"https:\/\/www.microsoft.com\/en-us\/research\/project\/uniprompt-a-structured-approach-to-prompt-optimization\/","title":{"rendered":"UniPrompt: A Structured Approach to Prompt Optimization"},"content":{"rendered":"
\n\t
\n\t\t
\n\t\t\t\t\t<\/div>\n\t\t\n\t\t
\n\t\t\t\n\t\t\t
\n\t\t\t\t\n\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\n\n

UniPrompt: A Structured Approach to Prompt Optimization<\/h1>\n\n\n\n

Algorithm to generate complex LLM prompts from scratch<\/p>\n\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t<\/div>\n\t\t<\/div>\n\t<\/div>\n<\/section>\n\n\n\n\n\n

Given a task in the form of a basic description and its training examples, prompt optimization is the problem of synthesizing the given information into a text prompt for a large language model (LLM). Humans solve this problem by also considering the different facets that define a task (e.g., counter-examples, explanations, analogies) and including them in the prompt. However, it is unclear whether existing algorithmic approaches, based on iteratively editing a given prompt or automatically selecting a few in-context examples, can cover the multiple facets required to solve a complex task. <\/p>\n\n\n\n

<\/p>\n\n\n\n

In this work, we view prompt optimization as that of learning multiple facets of a task from a set of training examples. We identify and exploit structure in the prompt optimization problem — first, we find that prompts can be broken down into loosely coupled semantic sections that have a relatively independent effect on the prompt’s performance; second, we cluster the input space and use clustered batches so that the optimization procedure can learn the different facets of a task across batches. The resulting algorithm, UniPrompt, consists of a generative model to generate initial candidates for each prompt section; and a feedback mechanism that aggregates suggested edits from multiple mini-batches into a conceptual description for the section. <\/p>\n\n\n\n

<\/p>\n\n\n\n

Empirical evaluation on multiple datasets and a real-world task show that prompts generated using UniPrompt obtain higher accuracy than human-tuned prompts and those from state-of-the-art methods. In particular, our algorithm can generate long, complex prompts that existing methods are unable to generate.<\/p>\n\n\n\n

\n

Code for Uniprompt will be released soon!<\/p>\n<\/blockquote>\n\n\n","protected":false},"excerpt":{"rendered":"

Algorithm to generate complex LLM prompts from scratch Given a task in the form of a basic description and its training examples, prompt optimization is the problem of synthesizing the given information into a text prompt for a large language model (LLM). Humans solve this problem by also considering the different facets that define a […]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"footnotes":""},"research-area":[13556],"msr-locale":[268875],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-1046808","msr-project","type-msr-project","status-publish","hentry","msr-research-area-artificial-intelligence","msr-locale-en_us","msr-archive-status-active"],"msr_project_start":"","related-publications":[],"related-downloads":[],"related-videos":[],"related-groups":[],"related-events":[],"related-opportunities":[],"related-posts":[],"related-articles":[],"tab-content":[],"slides":[],"related-researchers":[{"type":"user_nicename","display_name":"Amit Sharma","user_id":30997,"people_section":"Related people","alias":"amshar"},{"type":"user_nicename","display_name":"Nagarajan Natarajan","user_id":37311,"people_section":"Related people","alias":"nagarajn"},{"type":"guest","display_name":"Gurusha Juneja","user_id":1046811,"people_section":"Related people","alias":""}],"msr_research_lab":[199562],"msr_impact_theme":[],"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/1046808"}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-project"}],"version-history":[{"count":4,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/1046808\/revisions"}],"predecessor-version":[{"id":1046823,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/1046808\/revisions\/1046823"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1046808"}],"wp:term":[{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=1046808"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1046808"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1046808"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=1046808"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}