{"id":120706,"date":"2020-09-23T15:00:21","date_gmt":"2020-09-23T15:00:21","guid":{"rendered":"https:\/\/blogs.nvidia.com\/?p=46986"},"modified":"2020-09-23T15:00:21","modified_gmt":"2020-09-23T15:00:21","slug":"ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk","status":"publish","type":"post","link":"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/","title":{"rendered":"AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk"},"content":{"rendered":"

Paul Edwards is helping carry the age-old business of giving loans into the modern era of AI.<\/p>\n

Edwards started his career modeling animal behavior as a Ph.D. in numerical ecology. He left his lab coat behind to lead a group of data scientists at Scotiabank, based in Toronto, exploring how machine learning can improve predictions of credit risk.<\/p>\n

The team believes machine learning can both make the bank more profitable and help more people who deserve loans get them. They aim to share later this year some of their techniques in hopes of nudging the broader industry forward.<\/p>\n

Scorecards Evolve from Pencils to AI<\/b><\/h2>\n

The new tools are being applied to scorecards that date back to the 1950s when calculations were made with paper and pencil. Loan officers would rank applicants\u2019 answers to standard questions, and if the result crossed a set threshold on the scorecard, the bank could grant the loan.<\/p>\n

With the rise of computers, banks replaced physical scorecards with digital ones. Decades ago, they settled on a form of statistical modeling called a \u201cweight of evidence logistic regression\u201d that\u2019s widely used today.<\/p>\n

One of the great benefits of scorecards is they\u2019re clear. Banks can easily explain their lending criteria to customers and regulators. That\u2019s why in the field of credit risk, the scorecard is the gold standard for explainable models.<\/p>\n

\u201cWe could make machine-learning models that are bigger, more complex and more accurate than a scorecard, but somewhere they would cross a line and be too big for me to explain to my boss or a regulator,\u201d said Edwards.<\/p>\n

Machine Learning Models Save Millions<\/b><\/h2>\n

So, the team looked for fresh ways to build scorecards with machine learning and found a technique called boosting.<\/p>\n

They started with a single question on a tiny scorecard, then added one question at a time. They stopped when adding another question would make the scorecard too complex to explain or wouldn\u2019t improve its performance.<\/p>\n

The results were no harder to explain than traditional weight-of-evidence models, but often were more accurate.<\/p>\n

\u201cWe\u2019ve used boosting to build a couple decision models and found a few percent improvement over weight of evidence. A few percent at the scale of all the bank\u2019s applicants means millions of dollars,\u201d he said.<\/p>\n

XGBoost Upgraded to Accelerate Scorecards<\/h2>\n

Edwards\u2019 team understood the potential to accelerate boosting models because they had been using a popular library called XGBoost on an NVIDIA DGX system<\/a>. The GPU-accelerated code was very fast, but lacked a feature required to generate scorecards, a key tool they needed to keep their models simple.<\/p>\n

Griffin Lacey, a senior data scientist at NVIDIA, worked with his colleagues to identify and add the feature. It\u2019s now part of XGBoost in\u00a0RAPIDS<\/a>, a suite of open-source software libraries for running data science on GPUs.<\/p>\n

As a result, the bank can now generate scorecards 6x faster using a single GPU compared to what used to require 24 CPUs, setting a new benchmark for the bank. \u201cIt ended up being a fairly easy fix, but we could have never done it ourselves,\u201d said Edwards.<\/p>\n

GPUs speed up calculating digital scorecards and help the bank lift their accuracy while maintaining the models\u2019 explainability. \u201cWhen our models are more accurate people who are deserving of credit get the credit they need,\u201d said Edwards.<\/p>\n

Riding RAPIDS to the AI Age<\/b><\/h2>\n

Looking ahead, Edwards wants to leverage advances from the last few decades of machine learning to refresh the world of scorecards. For example, his team is working with NVIDIA to build a suite of Python tools for scorecards with features that will be familiar to today\u2019s data scientists.<\/p>\n

\u201cThe NVIDIA team is helping us pull RAPIDS tools into our workflow for developing scorecards, adding modern amenities like Python support, hyperparameter tuning and GPU acceleration,\u201d Edwards said. \u201cWe think in six months we could have example code and recipes to share,\u201d he added.<\/p>\n

With such tools, banks could modernize and accelerate the workflow for building scorecards, eliminating the current practice of manually tweaking and testing their parameters. For example, with GPU-accelerated hyperparameter tuning, a developer can let a computer test 100,000 model parameters while she is having her lunch.<\/p>\n

With a much bigger pool to choose from, banks could select scorecards for their accuracy, simplicity, stability or a balance of all these factors. This helps banks ensure their lending decisions are clear and reliable and that good customers get the loans they need.<\/p>\n

Digging into Deep Learning<\/b><\/h2>\n

Data scientists at Scotiabank use their DGX system to handle multiple experiments simultaneously. They tune scorecards, run XGBoost and refine deep-learning models. \u201cThat\u2019s really improved our workflow,\u201d said Edwards.<\/p>\n

\u201cIn a way, the best thing we got from buying that system was all the support we got afterwards,\u201d he added, noting new and upcoming RAPIDS features.<\/p>\n

Longer term, the team is exploring use of deep learning to more quickly identify customer needs. An experimental model for calculating credit risk already showed a 20 percent performance improvement over the best scorecard, thanks to deep learning.<\/p>\n

In addition, an emerging class of generative models can create synthetic datasets that mimic real bank data but contain no information specific to customers. That may open a door to collaborations that speed the pace of innovation.<\/p>\n

The work of Edwards\u2019 team reflects the growing interest and adoption of AI in banking.<\/p>\n

\u201cLast year, an annual survey<\/a> of credit risk departments showed every participating bank was at least exploring machine learning and many were using it day-to-day,\u201d Edwards said.<\/p>\n<\/div>

The post AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk<\/a> appeared first on The Official NVIDIA Blog<\/a>.<\/p>\n

\n<\/a> <\/a>\n<\/div>\"\"\/","protected":false},"excerpt":{"rendered":"

Paul Edwards is helping carry the age-old business of giving loans into the modern era of AI. Edwards started his career modeling animal behavior as a Ph.D. in numerical ecology. He left his lab coat behind to lead a group of data scientists at Scotiabank, based in Toronto, exploring how machine learning can improve predictions Read article ><\/a><\/p>\n

The post AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk<\/a> appeared first on The Official NVIDIA Blog<\/a>.<\/p>\n","protected":false},"author":303,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_mi_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"footnotes":""},"categories":[343],"tags":[5901,4223,9935,2181,4233,2043,6031,13218],"yoast_head":"\nAI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk - WebDomino.NET<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk - WebDomino.NET\" \/>\n<meta property=\"og:description\" content=\"Paul Edwards is helping carry the age-old business of giving loans into the modern era of AI. Edwards started his career modeling animal behavior as a Ph.D. in numerical ecology. He left his lab coat behind to lead a group of data scientists at Scotiabank, based in Toronto, exploring how machine learning can improve predictions Read article > The post AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk appeared first on The Official NVIDIA Blog.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/\" \/>\n<meta property=\"og:site_name\" content=\"WebDomino.NET\" \/>\n<meta property=\"article:published_time\" content=\"2020-09-23T15:00:21+00:00\" \/>\n<meta property=\"og:image\" content=\"http:\/\/feeds.feedburner.com\/~ff\/nvidiablog?d=yIl2AUoC8zA\" \/>\n<meta name=\"author\" content=\"John Ashley\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"John Ashley\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"5 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/\",\"url\":\"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/\",\"name\":\"AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk - WebDomino.NET\",\"isPartOf\":{\"@id\":\"http:\/\/webdomino.net\/#website\"},\"datePublished\":\"2020-09-23T15:00:21+00:00\",\"dateModified\":\"2020-09-23T15:00:21+00:00\",\"author\":{\"@id\":\"http:\/\/webdomino.net\/#\/schema\/person\/35d60a11a6a7da9981e193a7a4c9135a\"},\"breadcrumb\":{\"@id\":\"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"http:\/\/webdomino.net\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk\"}]},{\"@type\":\"WebSite\",\"@id\":\"http:\/\/webdomino.net\/#website\",\"url\":\"http:\/\/webdomino.net\/\",\"name\":\"WebDomino.NET\",\"description\":\"Global NEWS WordPress site\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"http:\/\/webdomino.net\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"http:\/\/webdomino.net\/#\/schema\/person\/35d60a11a6a7da9981e193a7a4c9135a\",\"name\":\"John Ashley\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"http:\/\/webdomino.net\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/f3b3362701a683f85772a7a345907a9b?s=96&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/f3b3362701a683f85772a7a345907a9b?s=96&d=mm&r=g\",\"caption\":\"John Ashley\"},\"sameAs\":[\"https:\/\/blogs.nvidia.com\"],\"url\":\"https:\/\/webdomino.net\/index.php\/author\/john-ashley\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk - WebDomino.NET","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/","og_locale":"en_US","og_type":"article","og_title":"AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk - WebDomino.NET","og_description":"Paul Edwards is helping carry the age-old business of giving loans into the modern era of AI. Edwards started his career modeling animal behavior as a Ph.D. in numerical ecology. He left his lab coat behind to lead a group of data scientists at Scotiabank, based in Toronto, exploring how machine learning can improve predictions Read article > The post AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk appeared first on The Official NVIDIA Blog.","og_url":"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/","og_site_name":"WebDomino.NET","article_published_time":"2020-09-23T15:00:21+00:00","og_image":[{"url":"http:\/\/feeds.feedburner.com\/~ff\/nvidiablog?d=yIl2AUoC8zA"}],"author":"John Ashley","twitter_misc":{"Written by":"John Ashley","Est. reading time":"5 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/","url":"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/","name":"AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk - WebDomino.NET","isPartOf":{"@id":"http:\/\/webdomino.net\/#website"},"datePublished":"2020-09-23T15:00:21+00:00","dateModified":"2020-09-23T15:00:21+00:00","author":{"@id":"http:\/\/webdomino.net\/#\/schema\/person\/35d60a11a6a7da9981e193a7a4c9135a"},"breadcrumb":{"@id":"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/webdomino.net\/index.php\/nvidia\/ai-scorekeeper-scotiabank-sharpens-the-pencil-in-credit-risk\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"http:\/\/webdomino.net\/"},{"@type":"ListItem","position":2,"name":"AI Scorekeeper: Scotiabank Sharpens the Pencil in Credit Risk"}]},{"@type":"WebSite","@id":"http:\/\/webdomino.net\/#website","url":"http:\/\/webdomino.net\/","name":"WebDomino.NET","description":"Global NEWS WordPress site","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"http:\/\/webdomino.net\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"en-US"},{"@type":"Person","@id":"http:\/\/webdomino.net\/#\/schema\/person\/35d60a11a6a7da9981e193a7a4c9135a","name":"John Ashley","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"http:\/\/webdomino.net\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/f3b3362701a683f85772a7a345907a9b?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/f3b3362701a683f85772a7a345907a9b?s=96&d=mm&r=g","caption":"John Ashley"},"sameAs":["https:\/\/blogs.nvidia.com"],"url":"https:\/\/webdomino.net\/index.php\/author\/john-ashley\/"}]}},"_links":{"self":[{"href":"https:\/\/webdomino.net\/index.php\/wp-json\/wp\/v2\/posts\/120706"}],"collection":[{"href":"https:\/\/webdomino.net\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/webdomino.net\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/webdomino.net\/index.php\/wp-json\/wp\/v2\/users\/303"}],"replies":[{"embeddable":true,"href":"https:\/\/webdomino.net\/index.php\/wp-json\/wp\/v2\/comments?post=120706"}],"version-history":[{"count":1,"href":"https:\/\/webdomino.net\/index.php\/wp-json\/wp\/v2\/posts\/120706\/revisions"}],"predecessor-version":[{"id":120707,"href":"https:\/\/webdomino.net\/index.php\/wp-json\/wp\/v2\/posts\/120706\/revisions\/120707"}],"wp:attachment":[{"href":"https:\/\/webdomino.net\/index.php\/wp-json\/wp\/v2\/media?parent=120706"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/webdomino.net\/index.php\/wp-json\/wp\/v2\/categories?post=120706"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/webdomino.net\/index.php\/wp-json\/wp\/v2\/tags?post=120706"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}