The Food Security Portal (FSP), facilitated by IFPRI, aims to provide improved food security for the world’s poor and increased resilience of global food systems against food and financial crisis. The project brings together international, regional, and country-level data, news, and research aimed at meeting countries’ immediate food security needs, as well as building long-term global food security. The FSP is designed to pool information in structured ways to ensure data quality, timeliness, and relevance, as well as the opportunity for collaboration among policymakers, development professionals, and researchers. The Food Security Portal (FSP), together with the Africa South of the Sahara Food Security Portal (FSP-SSA) sub-portals, available in English and in French, feature global food price data and price volatility measurements for major agricultural commodities.
The Food Security Portal’s Media Analysis System employs advanced linguistic and semantic object network-mapping algorithms to analyze the relationships between key terms found in media articles that may impact commodity price volatility. The system provides an intuitive interface designed to help users easily identify sentences that mention price movements (up or down) for each commodity. The system is also used to identify the media mentions of the occurrence of conflict situations worldwide. Daily, global food- and commodity-related and conflict-related news articles are loaded into the Food Security Media Analysis System. The system then updates the news article database and mines the complete corpus to generate up-to-date daily media analysis of factors that may influence commodity price volatility and food security. This analysis is performed using the General Architecture for Text Engineering (GATE), a powerful suite of tools for natural language processing and text mining.
The Food Security Portal seeks a qualified consultant or organization to upgrade the existing media analysis system by incorporating large language model (GPT-4) technology. This involves restructuring the current system to better adapt to the advanced capabilities of GPT-4. It also involves addressing limitations in the current analysis including: (a) duplicate news articles with differing titles but identical content and (b) erroneous information extraction from advertisements on the same page.
Deliverable
Upgraded Media Analysis Model – This involves restructuring the current system to incorporate GPT-4 technology and address known limitations in the current analysis including: (a) duplicate news articles with differing titles but identical content and (b) erroneous information extraction from advertisements on the same page.
Duration
This consultancy will take place from February 2025-April 2025.
Required qualifications of the consultant / team leader
Required Application Documents
Please include the following in your application:
In the event of any questions, please contact IFPRI-FSP@cgiar.org.
Application deadline January 24, 2025
Level of Education: Bachelor Degree
Work Hours: 8
Experience in Months: No requirements
Notifications