
Computer vision is of great importance to financial services, with applications including OCR, object detection and recognition, video analysis, and more. At Borealis AI, we work on problems in financial services. Often, we find that the research we do and the machine learning models we use are similar to those in computer vision. For example:
- Machine learning model architectures that are used in object recognition, activity recognition, or motion analysis/tracking, such as transformers, recurrent neural networks, or other deep models are often adapted to modeling financial data.
- Video forecasting, autonomous driving prediction, and other related work can be used to solve financial time-series prediction problems, such as transaction prediction, or stock price prediction.
- Graph-based methods commonly used to model dependency between various objects and agents can be utilized to do model interaction among customers.
- Representation learning techniques, including self-supervised learning approaches, are adapted to the type of data we deal with in finance.
- CV methods such as online learning, learning from new data streams, and avoiding catastrophic forgetting are also widely applicable in finance.
- We work on multi-modal data — combining various sources of client interactions (mobile app, transactions, etc.), and utilize techniques similar to those in vision+language or vision+other modalities lines of research.
- We work in domains with structure (e.g., multiple stocks, sectors they are organized by), using the techniques that are related to those in scene understanding.
- Explainable AI, fairness, accountability, privacy, transparency, and ethics form a core part of computer vision. At Borealis AI, deep integrity and building responsibly sit at the core of the work that we do that impacts millions of people’s lives and financial wellbeing.
- In addition, the technology advances from CVPR literature can usually be extracted from specific tasks; the underneath models and optimization techniques can also be widely adopted to financial applications.
Borealis AI conducts research and builds products for financial services
We are part of RBC, one of the largest banks in the world. At Borealis, we publish impactful research and build products for millions of people, helping to improve their financial lives. As part of RBC, we have access to data, problems, and domain expertise to drive our research.
We organize our research under three main North Star focus areas that guide our work, each typically formulated as a challenging problem we are trying to solve. Our three main areas of focus include the following core scientific problems:
- Asynchronous Temporal Models (ATOM). We build machine learning models capable of making inferences from partially-observed, multi-source, asynchronous temporal data. These are the types of data commonly found in banking applications — from various types of transaction data to client interactions with our banking services.
- Non-Cooperative Learning in Competing Markets (Photon). We build models for Capital Markets data, where challenges include low signal to noise ratio, structured prediction, and game theoretic impacts from decisions we make.
- Causal Machine Learning from Observational Data (Causmos). Our machine learning products drive decision making, so ultimately, we need models that can answer causal questions and generalize out of distribution.
Responsible AI is at the core of our work at Borealis AI
AI permeates our daily lives, and ensuring it is being developed and used in a responsible and ethical way has become a top priority. In finance, one of the most regulated industries on the planet, it is the only we build. Borealis AI’s research covers some of the key topics in this area including adversarial robustness, explainability, and fairness.
Cutting-edge research requires diversity of perspectives, tenacity, and creative thinking to fundamentally advance what is possible in Machine Learning
In finance, researchers with backgrounds across artificial intelligence including computer vision, machine learning, and natural language processing, and with PhDs in computer science, physics, computational finance, mathematics and more, often find the opportunity and resources to do impactful work. At Borealis AI, researchers get to work closely with engineers, product, and business experts to bring their research and prototypes to life, advancing what’s possible in ML and shaping the future of financial services.