Rural populations disproportionately lack access to improved water supplies. We evaluated a novel scheme that employed community-based sales agents to disseminate the Safe Water System (SWS)--a household-level water chlorination and safe storage intervention--in rural Madagascar. Respondents from 242 households in 4 villages were interviewed; all used surface water for drinking water. Respondents from 239 households (99%) had heard of Sûr'Eau, the SWS disinfectant; 226 (95%) reported having ever used Sûr'Eau, and 166 (73%) reported current use. Current Sûr'Eau use was confirmed in 54% of households. Community sales agents effectively motivated their neighbors to adopt a new health behavior that prevents diarrhea. Future work should focus on strategies for sustaining SWS use, factors that motivate community-based sales agents to promote SWS, and the feasibility of scaling up this approach.

When modernization and other changes demand workforce reskilling, employers often turn to local colleges for training programs. Doing so can be a frustrating experience. HR and talent professionals have difficulty identifying and communicating requirements, especially for new jobs and roles, while college continuing education (CE) and professional development offices have difficulty understanding and responding to company needs. This article describes an NSF Convergence Accelerator project called SkillSync™ in which multiple forms of AI are used to address this specific problem and provide national efforts (e.g., the US Chamber of Commerce Talent Pipeline Management initiative) with skills data and skills alignment services. Skillsync uses variations on the Siamese Multi‐depth Transformer‐based Hierarchical Encoder (SMITH) and other natural language understanding methods to map job descriptions and course information to skills taxonomies, uses machine‐learned models to align skills needs with learning outcomes and training, and incorporates an intelligent coach based on Georgia Tech's Jill Watson “virtual teaching assistant” to answer questions about Skillsync's vocabulary, functionality, and process. This article describes these AI methods, how these methods are used in Skillsync, and the challenges involved.

No abstract

No abstract

When modernization and other changes demand workforce reskilling, employers often turn to local colleges for training programs. Doing so can be a frustrating experience. HR and talent professionals have difficulty identifying and communicating requirements, especially for new jobs and roles, while college continuing education (CE) and professional development offices have difficulty understanding and responding to company needs. This article describes an NSF Convergence Accelerator project called SkillSync™ in which multiple forms of AI are used to address this specific problem and provide national efforts (e.g., the US Chamber of Commerce Talent Pipeline Management initiative) with skills data and skills alignment services. Skillsync uses variations on the Siamese Multi-depth Transformer-based Hierarchical Encoder (SMITH) and other natural language understanding methods to map job descriptions and course information to skills taxonomies, uses machine-learned models to align skills needs with learning outcomes and training, and incorporates an intelligent coach based on Georgia Tech's Jill Watson “virtual teaching assistant” to answer questions about Skillsync's vocabulary, functionality, and process. This article describes these AI methods, how these methods are used in Skillsync, and the challenges involved.