Machine Learning Helps Banks and Buyers Complete Real Estate Deals
The home buying process can feel like an obstacle course – finding the perfect place, writing an offer and, the biggest hurdle of all, getting a mortgage.
San Francisco-based real estate tech company Doma is helping would-be homeowners navigate this hurdle faster with the support of AI. Its machine learning models accelerate properties through the title search, underwriting and closing processes, enabling home transactions to be completed up to 15% faster.
“There’s a lot of paperwork involved in this process,” said Brian Holligan, director of data science at Doma. “The more successful we are in using machine learning to identify different types of documents and extract relevant information, the faster and more transparent the process can be. »
Doma uses machine learning to identify different types of real estate documents and extract information from these files. He also develops natural language understanding models to help everyone involved in a real estate transaction – from loan officers to real estate agents to home buyers – to quickly interpret the many requests and inquiries that typically occur. during the process.
Since its inception in 2016, Doma has accelerated over 100,000 real estate transactions using machine learning.
The company uses machine learning models – both transformer-based NLP tools and convolutional neural networks for computer vision – which leverage NVIDIA V100 Tensor Core GPUs through Microsoft Azure for model training .
“Working with a remote team is great to have the flexibility of GPUs in the cloud,” said Keesha Erickson, data scientist at Doma. “We can run the right size machines for the project or task at hand. If there is a larger scale model with a longer runtime, we can grab GPUs that are suitable for the time constraints we are under.
Doma Machine Intelligence dives into real estate documents
Once a seller and a buyer agree on the purchase price of a home, they enter into a contract. But usually a few weeks pass before the keys change hands – a period known as escrow. During this process, the buyer’s mortgage is finalized and a title company investigates home ownership, balance charges and tax history.
Doma’s GPU-accelerated machine learning models speed up the title review process by analyzing property records and mortgage records to help identify risks that could disrupt the transaction.
Like other fields such as drug discovery or architecture, the real estate industry has jargon that a general NLP model may not be able to interpret. To adapt its AI to this jargon, Doma is refining a suite of models, including BERT-based models, using a body of proprietary real estate data.
Doma’s technology also uses computer vision models to analyze older real estate documents. Many records come from county courthouses and clerks’ offices – and depending on the age of the house, the records may be scans of incredibly poor quality paper documents dating back decades.
Doma machine learning engineer Juhi Chandalia, who works on machine learning models for this type of document processing, found that using NVIDIA GPUs for inference reduced machine analysis time. team by 4, less than a minute.
“I’ve already started training models on CPUs instead of GPUs, and realized it would take weeks,” Chandalia said. “My team relies on NVIDIA GPUs because otherwise, by the time we finish training and testing our machine learning models, they would be obsolete.”
Doma offers application programming interfaces, prebuilt integrations and custom versions of its platform to its lender partners. The company has partnered with major mortgage lenders in the United States, including Chase, Homepoint Financial, PennyMac and Sierra Pacific Mortgage, to expedite the mortgage transaction and refinance process.
The company is also bringing some of its machine learning tools to individuals – real estate agents, buyers and sellers – to further streamline the complex process for all parties in a real estate transaction.
Learn how AI can help predict mortgage defaults, improve credit risk management, and power the banks of the future.
Subscribe to the latest NVIDIA financial services news.