Will Gaviria Rojas - Using AI to Bring Structure to Unstructured Data
Dec 28, 2023
auto_awesome
Will Gaviria Rojas, co-founder of CoactiveAI, and Joe Reis discuss the exciting world of unstructured data and AI, including fine-grain searching and image recognition. They explore the concept of data-centric AI, the shift towards multimodal models, and the challenges of working with massive amounts of data. The rapid generation of data sets and the future of AI-generated content are also discussed, along with the significance of expertise in working with unstructured data.
CoactiveAI aims to merge AI and data systems to analyze and extract insights from unstructured visual data.
Coactive prioritizes creating user-friendly interfaces to enable easy search, filtering, and understanding of visual data using AI-powered analytics.
Coactive recognizes the exponential growth of unstructured data and aims to bridge the gap between storage and users, merging data systems and AI algorithms for effective management and analysis.
Deep dives
Coactive's mission to help research and analyze visual data
Coactive, founded by Boa Gari De Rojas, focuses on helping users research, filter, and analyze visual data. They aim to address the growing need for understanding and making sense of unstructured data such as images, audio, and video. With the increasing accumulation of visual data by enterprises, Coactive recognized the valuable potential within this data. They aim to merge AI and data systems to provide scalable solutions for analyzing and extracting insights from visual data.
The need for accessible interfaces in visual data analytics
Coactive emphasizes the importance of creating user-friendly interfaces for working with visual data. They aim to make their platform accessible not only to ML developers and software engineers but also to laypersons. By creating easy-to-use interfaces, Coactive enables users to search, filter, and make sense of their visual data using AI-powered analytics. They draw inspiration from successful AI tools that offer simple interfaces, enabling users with varying levels of technical expertise to interact with the platform.
Challenges and opportunities in managing unstructured data
Coactive identified the challenge posed by the exponential growth of unstructured data, including images, video, text, and more. Many enterprises accumulate vast amounts of unstructured data without effectively utilizing it. Coactive recognized the value in bridging the gap between the storage layer and the users, providing simple interfaces to efficiently search, filter, and understand unstructured data. Their goal is to merge data systems and AI algorithms to address the need for scalable, effective management and analysis of unstructured data.
The emerging field of multimodal data and AI
Coactive acknowledges the significant potential of multimodal data, combining images, video, text, and audio. They foresee a future where AI-powered tools can understand and analyze data spanning these different modalities. Coactive emphasizes the upcoming challenges and opportunities in handling large-scale multimodal data. Additionally, they highlight the importance of considering specific nuances and properties associated with different types of visual data, such as TikTok videos, Zoom calls, surveillance footage, and cinematic video. Coactive aims to play a role in facilitating the development of tools and infrastructure for effective multimodal data analysis.
The future of data-centric AI and benchmarking
Coactive discusses the importance of data-centric AI and its impact on the field of benchmarking. They believe in shifting the focus from solely optimizing models to also optimizing data, as data quality and quantity play a crucial role in AI performance. Coactive has initiated the DataPerf benchmark, which aims to evaluate the performance of data-centric AI systems. They propose examining the minimum dataset sizes needed for accurate training, as well as creating challenging test sets. By focusing on data quality and performance, they aim to drive advancements and improvements in AI systems.
Will Gaviria Rojas (co-founder of CoactiveAI) and I sit in a hotel bar and chat about the amazing world of unstructured data and AI. We chat about searching unstructured datasets (without metadata), the state of AI, and much more. Will is one of my favorite people in general, so I really enjoyed our chat. Enjoy!
Get the Snipd podcast app
Unlock the knowledge in podcasts with the podcast player of the future.
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode
Save any moment
Hear something you like? Tap your headphones to save it with AI-generated key takeaways
Share & Export
Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode