This is Machine Learning at Capital One

From random forests to causal models, explore how we use machine learning for better banking

We’re changing the financial services field by creating real-time, intelligent, automated customer experiences using AI & ML. From informing customers about unusual charges to answering their questions in real time, our work in AI & ML is bringing humanity and simplicity to banking.

Let us show you how.

Improving Eno’s Virtual Card Numbers with Edge Machine Learning

The Eno browser extension uses machine learning on device to read the DOM & pixel coordinates for a more accurate VCN experience. Read how edge machine learning helped us build a faster, more secure customer experience.

Using Machine Learning to Automate Detection, Diagnosis & Remediation of Mobile App Failure

Learn how we’re using machine learning to enable the automation of incident management on the Capital One Mobile app and beyond. For more on how machine learning and DevOps go hand in hand at Capital One click here.

How Machine Learning Can Help Fight Money Laundering

Can a machine learning model help identify suspicious account activity and better support anti-money laundering teams? Read how machine learning is supercharging our AML efforts here.

Read More About Our Work in AI & ML Here

• How to Fine-Tune Sentence-BERT for Question Answering — A simple introductory tutorial on how to use the sentence-transformers library to fine-tune Sentence-BERT for question matching purposes.
• Dynamic Customer Embeddings & Understanding Customer Intent — Representation learning and temporal sequence modeling remain essential building blocks for providing teams with the meaningful features needed to build effective, personalized systems for customer experiences.
• 10 Common Machine Learning Mistakes and How to Avoid Them — Avoiding machine learning mistakes can be a challenge, but is especially important when working on complex projects at scale.
• A Modern Dilemma: When to Use Rules vs. Machine Learning — Understanding the strengths of both rules engines and machine learning can help identify the right solution for a problem. Especially in cases where using both together can lead to maximum value.
• Custom Machine Learning Estimators at Scale on Dask & RAPIDS — This article discusses how to leverage the scikit-learn library’s API to add customizations that can minimize code, reduce maintenance, facilitate reuse, and provide the ability to scale with technologies such as Dask and RAPIDS.
• How to Control Your XGBoost Model — Pruning, regularization, and early stopping are all important tools that control the complexity of XGBoost models. But they also come with many quirks that can lead to unintuitive behavior. By learning more about what each parameter in XGBoost does you can build models that are smaller and less prone to overfitting the data.
• Innovative Curriculum Equips Teachers and Students — How Capital One’s Bot Camp program reached 80,000 participants, helping them gain valuable technology skills for the future.
• An Introduction to 5 Must-Know Machine Learning Algorithms — A five-part series that presents common machine learning algorithms in a way that balances technical info with approachable examples.
• End-to-End Models for Complex AI Tasks — Deep neural networks trained end-to-end can outperform classical pipeline-based systems — but they pose some challenges.
• A Leader’s Journey To Connect Women in Tech — Learn why Cat, a leader in Machine Learning at Capital One, is passionate about women’s empowerment and creating more opportunities for diverse tech candidates.

DISCLOSURE STATEMENT: © 2021 Capital One. Opinions are those of the individual author. Unless noted otherwise in this post, Capital One is not affiliated with, nor endorsed by, any of the companies mentioned. All trademarks and other intellectual property used or displayed are property of their respective owners.