This is part 2 of the series answering one of the most popular questions on Artificial Intelligence (AI). What is the reasoning behind the claims of the “black box problem” by data scientists when it comes to machine learning and AI? Read Staying ahead of the game with AI (Part 1): Unlock the black box. 

The explainability of artificial intelligence or machine learning (AI/ML) systems outcomes often becomes a concern when used in the financial sector. ML models are often referred to as black boxes because they are not directly explainable by the user (Guidotti and others 2019). 

Explainability is a complex and multifaceted issue. There are several reasons why ML models are frequently considered to be black boxes:

1. they are complicated and cannot be easily interpreted;
2. their input signals/source might not be known;
3. they are an ensemble of models rather than a single independent model; and
4. the prediction reasoning is unknown.

Without unboxing the “Black Box”, the trust and appropriateness of ML credit decisions are difficult for industry players to accept. Robustness of the model may be undermined due to exposure to vulnerabilities – such as biased data, unsuitable modeling techniques, or incorrect credit decision making.

AI comes with a promise to help businesses fulfill their business goals. But without a correct demonstration of AI and how it is programmed, people might have lesser trust for AI-powered systems.

This is where Juris Mindcraft succeeds –  it is an explainable AI, which means it can provide an explanation behind every decision reached. Juris Mindcraft allows users to:

• leverage on feature importance comparison which displays the importance of each of the features or attributes,
• use statistics or past data to support the reason behind prediction, and
• see the exact order of the top predictors at a customer level, giving them confidence in underlying signals or analysis driving model performance.

Juris Mindcraft is JurisTech’s very own proprietary AI, an automated Machine Learning (autoML) and artificial intelligence (AI) platform. It uses advanced machine learning (ML) techniques to build powerful AI models. An effortless AI that enables enterprises especially banks and financial institutions to make intelligent business decisions and gain insights to solve real-world problems.

Thanks to Juris Mindcraft, we can now provide an explanation behind every decision reached. For the other “Black boxed” AI’s in the market, we usually do not have the idea of how the dataset turned into information. However, Juris Mindcraft can evaluate and interpret the model with all types of explainability methods such as evaluation metrics and confusion matrix and etc. or by the model behaviours. This can improve human readability and understand why an AI/ML model is giving specific results. On top of that, you can debug unbias or odd behaviours from the model to avoid discrimination and sociteal bias. Furthermore, Juris Mindcraft has out-of-the-box capabilities such as machine learning, cognitive behavioral scoring, and continuous and autonomous self-learning to improve accuracy. Its capabilities stretch in all areas of the banking industry and it can be applied to other fields such as onboarding customers, product recommendations, and collections.

Curious to learn more about the capability of Juris Mindcraft, request a free demo now at contact@juristech.net. 

References:

• https://worldline.com/en/home/knowledgehub/blog/2021/january/ever-heard-of-the-aI-black-box-problem.html
• https://www.thinkautomation.com/bots-and-ai/the-ai-black-box-problem/
• https://ec.europa.eu/research-and-innovation/en/horizon-magazine/opening-black-box-artificial-intelligence
• https://bigcloud.global/the-difference-between-white-box-and-black-box-ai/
• https://towardsdatascience.com/explainable-ai-the-key-to-open-black-boxes-4ad09e04d791

About JurisTech

JurisTech (Juris Technologies) is a leading Malaysian-based fintech company, specialising in enterprise-class software solutions for banks, financial institutions, and telecommunications companies in Malaysia, Southeast Asia, and beyond.

The post Unboxing the “Black Box”: The Need for Explainability appeared first on JurisTech.

This is a series answering one of the most popular questions on Artificial Intelligence (AI). What is the reasoning behind the claims of the “black box problem” by data scientists when it comes to machine learning and AI?

Artificial Intelligence or Machine Learning (AI/ML) solutions in banking have matured and made major advances over the past decade. Today’s AI systems can perform well-defined tasks quite well – tasks that typically require human intelligence. The learning process takes the form of ML, which relies on mathematics, statistics, and decision theory.

• McKinsey & Company (2020) estimates the potential value of AI in the banking sector to reach $1 trillion.
• A recent survey of financial institutions (WEF 2020) shows that 77% of all respondents anticipate AI will be of high or very high overall importance to their businesses within two years.
• Bank of England (2020) and McKinsey & Company (2020) find that a considerable number of financial institutions expect AI/ML to play a bigger role after the COVID-19 pandemic.

Below are two of the most compelling reasons for banks and Fintechs to start embracing AI in lending/credit scoring. 

• More Accurate Credit Scoring

AI/ML predictive models can enhance the credit scoring process in the calculation of default and repayment risks. Based on research, ML reduces banks’ losses on delinquent customers by up to 25% (Khandani, Adlar, and Lo 2010). There is also potential for AI/ML to be used in commercial lending decisions for risk quantification of commercial borrowers. AI/ML reduces turnaround time and increases the efficiency of lending decisions.

• Taping on the Underserved Segment

There is also evidence that automated credit underwriting benefits the underserved segment resulting in higher accuracy in predicting defaults and higher approval rates (Gates, Perry, and Zorn 2002). AI/ML allows for more creative decisioning processes which harness alternative data such as social, business, location, and internet data, in combination with conventional data.

Even if a customer does not have or lacks credit history, AI/ML can generate a credit score by analysing the borrower’s digital footprint such as social media activities, bills payment history, and search engine activities.

To find out how Juris Mindcraft works to solve the black box problems of the above use cases, tune in to our next series! To read more about Juris Mindcraft, click here.

The post Staying ahead of the game with AI: Unlock the black box appeared first on JurisTech.