Data must drive every decision that modern business makes. However, most businesses have big blind spots. They don't know what's going on with their visual data.
Coactive is working to change it. Cody Coleman '13, Meng '15, and company founded by William Gaviria Rojas '13 has created an artificial intelligence-driven platform that allows you to understand data such as images, audio, and video to unlock new insights.
Coactive's platform instantly searches, organizes and analyzes unstructured visual content to help businesses make faster and better decisions.
“In the first big data revolution, businesses were better able to get value from structured data,” Coleman said, referring to data from tables and spreadsheets. “However, around 80-90% of the world's data is currently unstructured. In the next chapter of big data, companies will need to process data such as large images, video, audio, and more, and AI is an important part of unlocking its capabilities.”
Coactive already works with several large media and retailers to help you understand visual content without relying on manual sorting or tagging. This will help you get the right content for your users faster, remove explicit content from the platform, and reveal how certain content affects your behavior.
More broadly, founders believe that Coactive serves as an example of how AI can work more efficiently and solve new problems.
“The word Coactive means working together at the same time. That's our grand vision: helping humans and machines work together,” says Coleman. “We believe vision is more important than ever because AI can either separate us or connect us. We want to be an agent where Coactive draws us and gives us a new set of powers to humans.”
Give your computer vision
Coleman met Gaviria Rojas in the summer before he first ran the MIT Intermediate Edge Program year after year. Both major in electrical engineering and computer science, working to bring MIT OpenCourseware content to other projects to universities in Mexico.
“It was a great example of entrepreneurship,” Coleman recalls the OpenCourseware project. “It really empowered me to be responsible for business and software development, and then I started my own little web development business and decided to take it. [the MIT course] The founder's journey. ”
Coleman first explored the power of AI at MIT while working as a graduate researcher at the Digital Learning Office (now MIT Open Learning) to study how humans learn in MITX, which hosts large, open online courses created by MIT faculty and instructors.
“It was truly amazing to me to be able to democratize this transformational journey we had with MIT with digital learning, and to be able to apply AI and machine learning to create adaptive systems that will help not only understand how humans learn, but also provide a more personalized learning experience for people around the world,” Coleman said of Mitx. “It was also the first time I'd explored video content and applied AI.”
After MIT, Coleman went to Stanford University for his PhD, where he worked on lowering barriers to AI use. Through this research, he collaborated with companies such as Pinterest and Meta on AI and machine learning applications.
“It's where people could look at the future of AI and what they want to do with their content,” Coleman recalls. “We were looking at how big companies were using AI to drive business value, and we were looking at where Coactive's first spark came from.
Meanwhile, Gabilia Rojas Moved to the Bay Area in 2020 eBay data scientist. As part of the move, he had to help transport the couch, and Coleman was the lucky friend he called.
“In the car, we both noticed that we saw an explosion happening around data and AI,” says Gaviria Rojas. “At MIT, we saw people inventing technology that gained the front row seating of the big data revolution and unlocked value from that data. Cody and I realized they were about to explode in companies that collect huge amounts of data, but this time it was multimodal data like images, video, audio and text.
The platform the founders have continued to build – what Coleman calls the “AI operating system” is model agnosticism. This means that as the model continues to improve, the company can replace AI systems under the hood. Coactive's platform includes pre-built applications, such as business customers searching for content, generating metadata, extracting insights and doing analysis.
“Before AI, computers see the world through part-time jobs, while humans see the world through vision,” says Coleman. “Now with AI, machines will ultimately be able to see the world like we do, which will blur the world of digital and physics.”
Human Computer Interface Improvements
Reuters' image database provides millions of photos to journalists around the world. Before Coactive, the company relied on reporters to manually enter tags on each photo, so the appropriate images were displayed when journalists searched for a particular subject.
“People didn't add tags because it was incredibly slow and expensive to go through all these raw assets,” says Coleman. “So when I searched things, even if the relevant photos were in the database, the results were limited.”
Now, if a journalist on the Reuters website chooses “Enable AI Search,” Coactive can lift relevant content based on an AI system's understanding of each image and video details.
“We've significantly improved the quality of reporters' results, so we can tell a more accurate story than ever before,” Coleman says.
Reuters isn't just struggling to manage all of their content. Digital asset management is a huge component of many media and retail companies, and today it often relies on metadata manually to sort and search for its content.
Another co-customer is Fandom. It is one of the world's largest platforms for information on television, video games and movies, with over 300 million active users each month. Fandom uses Coactive to help you understand visual data from online communities and remove excessive gore and sexualized content.
“It took Fandom 24-48 hours to review new content,” says Coleman. “We can now use Coactive to codify community guidelines and generate finer grain information in an average of about 500 ms.”
In all use cases, the founder believes that it is Coactive to enable new paradigms in the way humans work on machines.
“Through the history of human computer interaction, we had to bend the keyboard and mouse and enter information in a way that machines could understand,” says Coleman. “Now, for the first time, we can speak naturally, share images and videos with AI. It allows us to understand that content. It is a fundamental change in the way we think about the interactions of human computers. The core vision of that change requires a new operating system and a new way of using content and AI.”
