ReddRadar
What is Carlos Toxtli's AI Portfolio?
This portfolio showcases the diverse Artificial Intelligence work of Carlos Toxtli, a Ph.D. in Computer Science and Assistant Professor at Clemson University. His focus lies in applying Human-Centered Artificial Intelligence to design tools that foster fair and frictionless interactions in the workplace, contributing to the Future of Work. The projects span various AI domains, integrating principles of transparency, fairness, and inclusion into socio-technical systems.
Drawing from experience at institutions like the United Nations, Google, Microsoft Research, and Snap Inc., the portfolio includes developments in Natural Language Processing for chatbots and feedback systems, Computer Vision for video analysis, Quantum Machine Learning for drug discovery, automated video editing, music generation, content moderation for gig economy platforms, identity validation, recommendation systems, and extensive work in Robotic Process Automation involving web crawlers and workflow design. It also features numerous teaching materials and publications in top academic venues.
Features
- Natural Language Processing (NLP): Development of chatbots (MATT, TaskBot, HolaGus, YBot Studio) for tasks like micro-tutoring, task delegation, and customer service.
- Computer Vision (CV): Implementation of tools like DeepStab for real-time video stabilization and DeepPiracy for video piracy detection using Deep Learning.
- Multimodal Analysis: Research framework (MultiAffect) for emotion recognition using video categorization and regression tasks.
- Quantum Machine Learning (QML): Advisory role in applying QML to drug discovery algorithms.
- Automatic Video Editing: Creation of AutomEditor for blooper recognition and automatic monologue video editing.
- Music Generation: Development of Hum2Song for multi-track polyphonic music generation from voice melodies using Neural Networks.
- Content Moderation: System (UnfairReviews) to detect inaccurate and unfair reviews for gig workers.
- Robotic Process Automation (RPA): Design of finite state machines for orchestrating web crawlers and development of numerous crawlers for data extraction.
- Recommendation Systems: Creation of ExpertTwin, an AI agent providing content recommendations to knowledge workers.
- Teaching Materials: Extensive collection of interactive notebooks covering AI, Machine Learning, Deep Learning, and RPA concepts.
Use Cases
- Developing chatbots for automated customer service or internal task management.
- Implementing AI for video analysis, stabilization, or content identification.
- Creating systems for fair content moderation, particularly in gig economies.
- Generating musical accompaniments automatically based on vocal input.
- Automating video editing processes by identifying and removing errors.
- Building recommendation engines tailored to knowledge workers.
- Automating data collection and process workflows using RPA and web crawlers.
- Utilizing AI for fraud detection in payment systems.
- Exploring Quantum Machine Learning applications, such as in drug discovery.
- Accessing educational resources and notebooks for learning AI and RPA.
