NetflixVideo annotator: building video classifiers using vision-language models and active learning
Efficiently build video classifiers with vision-language models and active learning using Video Annotator framework
MetaPVF: A novel metric for understanding AI systems’ vulnerability against SDCs in model parameters
Introducing PVF, a novel metric to measure vulnerability of AI systems to SDCs in model parameters and guide hardware system design
AWS MLDeploy a Slack gateway for Amazon Bedrock
Streamline collaboration in Slack with generative AI using Amazon Bedrock
AWS MLMaximize your Amazon Translate architecture using strategic caching layers
Optimize Amazon Translate architecture with caching layers for faster, cost-effective translations
Snorkel AIHow does the Snorkel Flow label model work?
Exploring the workings of the Snorkel Flow label model in the context of machine learning training datasets.
OpenAIAchieving 10x growth with agentic sales prospecting
Strategies for exponential growth through proactive sales prospecting
OpenAISurging developer productivity with custom GPTs
Maximize efficiency and output with personalized GPT models
Snorkel AIHow does the Snorkel Flow label model work?
Understanding the workings of the Snorkel Flow label model for efficient machine learning training dataset creation.
Apple MLMultimodal Large Language Models with Fusion Low Rank Adaptation for Device Directed Speech Detection
Technique for adapting multimodal large language models using Fusion Low Rank Adaptation for improved device-directed speech detection
Apple MLConformer-Based Speech Recognition on Extreme Edge-Computing Devices
Implementing Conformer-Based Speech Recognition efficiently on limited resource edge-computing devices for optimal performance
Apple MLComparative Analysis of Personalized Voice Activity Detection Systems: Assessing Real-World Effectiveness
Exploring the effectiveness of Personalized Voice Activity Detection systems through a comparative analysis
MIT AIMIT-Takeda Program wraps up with 16 publications, a patent, and nearly two dozen projects completed
MIT-Takeda Program concludes with multiple publications, a patent, and innovative AI applications in drug development and health care.