PharmaChat
An AI-driven agricultural question-answering application that allows farmers to ask questions in their native languages and returns relevant video clips or text answers, based on small language models and retrieval-augmented generation.
Timeline (last 90 days)
Project Gecko developed the PharmaChat app, which allows farmers to ask questions in their native language and returns the segment of a video that answers the question.
PharmaChat compresses domain knowledge into a locally deployable model by fine-tuning a small language model (SLM) to reduce infrastructure requirements.
The key technical paradigm behind PharmaChat-type applications is retrieval-augmented generation (RAG).
The project team conducted field tests in Kenya to observe farmers' acceptance of PharmaChat, their initial questions, and usage challenges.
PharmaChat added the MMCT Agent, based on an intelligent video understanding system, which can provide multimodal answers including video segments and specific chapters of community farming books.
All Facts (5)
Project Gecko developed the PharmaChat app, which allows farmers to ask questions in their native language and returns the segment of a video that answers the question.
50%UnverifiedPharmaChat compresses domain knowledge into a locally deployable model by fine-tuning a small language model (SLM) to reduce infrastructure requirements.
50%UnverifiedThe key technical paradigm behind PharmaChat-type applications is retrieval-augmented generation (RAG).
50%UnverifiedPharmaChat added the MMCT Agent, based on an intelligent video understanding system, which can provide multimodal answers including video segments and specific chapters of community farming books.
50%UnverifiedThe project team conducted field tests in Kenya to observe farmers' acceptance of PharmaChat, their initial questions, and usage challenges.
50%