How Mattermark uses Machine Learning
What is Machine Learning?
Machine learning is a computer's way of learning from examples
Machine Learning Usage
- Spam Filtering
- Search Engines
- Movie Recommendations
- Speech Recognition
- Medical Diagnosis
How does Mattermark use Machine Learning?
- Industry Tagging
- News Article subject attribution
- Similar Companies
- News Fact Extraction
- Semantic Search
Industry Tagging
News Attribution
- Customers want to keep track of news of the companies that the follow
- Mattermark crawls over a hundred sources of news daily
- Finding what company (if any) is the subject of a news article is challenging, since it could be one of 500k
News Fact Extraction
- When you read an article from TechCrunch about a company raising a Series A round, you absorb the knowledge
- We will do it automatically for 1000s of articles a day
Similar Companies
- Customers want easy access to comparable companies
- But 1M companies means there are 1 trillion potential relationships
- Impossible to do this manually, and difficult even to compare automatically
- Currently have ~300k companies with similar companies
What will Mattemark use ML for in the near-future?
Emerging Industries
- Build a model that keeps track of new companies whose descriptions don't seem to fit in our current industry taxonomy
- When a set of these companies cluster together with a critical mass, we can then have a human look at it and name the industry
Funding Prediction
- Customers would love to know which companies are going to be raising money in the next 6 months
- VCs/Investors want to a heads up for hot new deals
- Biz Dev clients want to know who to start building a relationship with
- Given all the historical data we have now (going back about a year), we can use that as training data very soon
Graph Search like querying
- Like Siri, but for business!
- Users should be able to type the question:
- "Companies in the Bay Area in the E-Commerce Industry with more than $10M in funding"
- This will be particularly useful for Mobile clients
Why is this important?
- Machine Learning allows us to do things that would be near-impossible for humans to do
- Machine Learning allows us to scalably do things that would take forever for humans to do
- This gives us a competitive advantage
- Many of our competitors use data analysts instead
- We can do more with fewer people
Who's doing ML at Mattermark?
We need more people! Refer anyone that you know that does Data Science/ML
Questions?