DSS NYC 2024 - where Finance meets AI and Machine Learning

By Konrad Budek

With guests from S&P, Fitch Ratings, and Deutsche Bank, the upcoming Data Science Salon 2024 conference will encompass a plethora of finance-related topics, including credit risk, navigating bias in datasets, and applications of Large Language Models, among others.

The data-rich (and generally rich) banking sector is expected to invest approximately 45.19 billion US dollars by the end of 2024. And it is only the beginning—the sector is expected to spend up to 97 billion US dollars on AI-based solutions in 2027 alone, nearly doubling its yearly expenditure in just three years.

With the dynamics so high and expenses so significant, it is no surprise that companies are looking for reliable and trustworthy sources of information about Artificial Intelligence (AI) and Machine Learning (ML). The upcoming Data Science Salon New York is a perfect example of an event where practitioners and experts share their insights and knowledge about applying AI- and ML-based solutions in financial institutions, including banks, insurers, and rating companies.

Beyond Financial Services: How to Leverage Financial and App Engagement Data to Build a Personalized Shopping Experience

Claudia Pereira Johnson, Head of Marketplace Search and Personalization at Nubank, will share her insights on leveraging financial and app engagement data to build a personalized shopping experience.

Her talk explores how integrating financial and app data can enhance online shopping, highlighting success stories in personalization. She will discuss how transaction histories, payment preferences, and in-app behaviors are used to understand consumer preferences.

“App Engagement is a sign of relevance and mindshare in the digital world, in the same way that brand awareness has been for traditional marketing and in particular for consumer packaged goods (CPG). When developing a digital product or content, we are no longer competing for space on shelves at the store, we are competing for customers' finite attention/eyeballs, while they have infinite alternatives” comments Claudia Pereira Johnson.

The expert will also address the role of data science in creating personalized recommendations and boosting user loyalty, emphasizing ethical data use and maintaining user trust. Real-world examples from Nubank’s Marketplace will illustrate how these strategies are engaging customers in Latin America. Attendees will gain practical strategies to exceed customer expectations and improve app experiences, with an open forum at the end for sharing ideas and visions for the future of personalized retail.

Navigating Credit Risk: Insights for Corporates and Financial Institutions

Prasad Tamminaina, Senior Quantitative Analyst from S&P Global, will deliver a talk about navigating credit risk for corporates and financial institutions. He will explore effective strategies for managing corporate credit risk using S&P Global Market Intelligence's Credit Analytics tools. Additionally, he will discuss how advanced analytics, including machine learning and deep learning, are transforming credit risk evaluation.

“As Financial institutions operate in one of the heavily regulated environments, they have more risks and challenges than any industries when implementing AI solutions. Financial data is highly sensitive, and AI solutions requires large volumes of data to train effectively. Ensuring data privacy and security throughout the AI lifecycle, from data collection to model deployment is critical “ comments Prasad Tamminaina.

“AI solutions may also introduce new operational risks, such as system failures, algorithmic errors, or cyber-attacks targeting AI infrastructure. Ensuring the resilience and reliability of AI systems through robust testing, monitoring, and incident response procedures is crucial” he adds.

Navigating Bias in AI: Promoting Ethical Decision-Making in Fintech

Sumedha Rai, Senior Data Scientist from Acorns Grow, will share her remarks on navigating bias in AI, how AI biases in fintech can shape our financial future, and what we're doing to create a fairer path.

“When it comes to financial decisions made using AI, discriminatory results produced by biased AI algorithms, can perpetuate existing inequalities and lead to unfair outcomes. For instance, biased models used in loan approval processes might unfairly deny credit to certain demographic groups based on historical data that reflects past prejudices, thereby intensifying financial exclusion. People from certain professions, income brackets, or age groups may receive higher mortgage rates, and these attributes may be one of the main reasons why their interest rates are higher than the rest” comments Sumedha Rai.

In the fast-paced world of fintech, hidden biases in AI algorithms can affect everyone. Join me to uncover these biases and their impact on fair access. Learn strategies to combat AI bias and explore ethical regulatory frameworks. This talk will provide actionable insights into building a more inclusive financial system through transparency, accountability, and ethical innovation.

“This kind of bias can result in certain populations being systematically disadvantaged, which not only affects individual financial opportunities but also has broader socio-economic implications. The detrimental effect of bias might not be limited to individuals. The use of biased AI systems can undermine trust in financial institutions. Customers and stakeholders who perceive decision-making processes as discriminatory may lose confidence in the institutions, leading to irreparable damage to their reputation, loss of current and future customers, and potential regulatory repercussions. It’s very important for financial institutions to recognize and address bias in their AI systems to maintain both fairness and trust” she adds.

Decoding LLMs: Challenges in Evaluation

Jayeeta Putatunda, Senior Data Scientist from Fitch Ratings, is going to share her remarks on evaluating LLMs. Large Language Models (LLMs) have revolutionized natural language processing, impacting fields from conversational AI to content generation. However, evaluating these complex models poses significant challenges.

According to the expert, there's a lack of standardized benchmarks to capture their diverse capabilities, and their black-box nature makes understanding decisions and biases difficult. This talk will explore effective evaluation metrics for LLMs and their alignment with real-world applications. As new architectures emerge, continuous and adaptable evaluation methods are essential.

She will also discuss the crucial role of open-source initiatives in developing standardized benchmarks and consistently evaluating LLM performance. Additionally, we'll review some open-source evaluation metrics and demonstrate their application using Kaggle data.

Exploring Opportunities and Challenges: Topological Data Analysis in Financial Discovery

Zach Golkhou, Director of AI and Data Science at J.P. Morgan, will talk about how the exponential growth of data in financial institutions offers great opportunities but also significant challenges. Traditional AI methods often struggle with the complexity and noise in large datasets. To address this, Topological Data Analysis (TDA) has emerged as a powerful tool due to its ability to uncover robust patterns despite variations in data.

According to the expert, TDA is especially valuable in finance, where interpretability and reliability are crucial. However, TDA's computational demands and the need for subsampling can limit its effectiveness. Recent advancements in Generative AI and tensor-based network algorithms are making TDA more practical and impactful. This talk will explore how to leverage topology in finance to gain deeper insights and improve data analysis beyond traditional methods.

Discussion Panels and Networking

The conference will also be enriched with a discussion panel about implementing and scaling AI/ML solutions in the financial sector. The panel will include experts from S&P Global, New York Life Insurance Company, and Capital One.

The conference will take place at S&P Global Headquarters in New York, NY, on June 18, 2024. The event will also be a great opportunity for networking and connecting with like-minded specialists in the field. You can order your tickets here.

Get the latest data science news and resources every Friday right to your inbox!