

Data Skeptic
Kyle Polich
The Data Skeptic Podcast features interviews and discussion of topics related to data science, statistics, machine learning, artificial intelligence and the like, all from the perspective of applying critical thinking and the scientific method to evaluate the veracity of claims and efficacy of approaches.
Episodes
Mentioned books

Jan 29, 2025 • 40min
Auditing LLMs and Twitter
In this insightful discussion, Erwan Le Merrer, a collaborator in graph theory and distributed systems, reveals how graph-based techniques can expose patterns in large language models and shadow banning on Twitter. He explains the application of epidemic models to examine shadow banning spread across user networks. The conversation also highlights the use of graph metrics to audit LLM outputs, the challenge of bias detection in AI, and innovative methodologies for understanding algorithmic behavior, shedding light on often-overlooked platform moderation practices.

7 snips
Jan 22, 2025 • 37min
Fraud Detection with Graphs
Šimon Mandlík, a PhD candidate specializing in machine learning for cybersecurity at the Czech Technical University, dives into the intriguing world of fraud detection using graph-based techniques. He explains how graphs can unveil malicious activities by analyzing relationships within vast datasets. The discussion highlights the advantages of his hierarchical multi-instance learning method over traditional approaches, tackling challenges like scalability and heterogeneous graphs. Mandlík emphasizes the 'locality assumption' in fraud detection, resulting in faster and more accurate outcomes.

23 snips
Jan 15, 2025 • 38min
Optimizing Supply Chains with GNN
Thibaut Vidal, a professor at Polytechnique Montreal, specializes in using advanced algorithms and machine learning for supply chain optimization. He discusses how graph-based methods can revolutionize logistics by improving routing and decision-making. The conversation highlights the effectiveness of Graph Neural Networks in predicting delivery costs for companies like UPS and Amazon. Thibaut also emphasizes the potential of these advanced techniques to cut costs, enhance efficiency, and create better working conditions through smarter route planning.

15 snips
Jan 10, 2025 • 48min
The Mystery Behind Large Graphs
David Tench, a Grace Hopper postdoctoral fellow at Lawrence Berkeley National Labs, specializes in scalable graph algorithms. He discusses how his techniques enable real-time analysis of massive datasets while reducing storage needs. David challenges the idea that large graphs are typically sparse, suggesting a potential bias in data analysis processes. He emphasizes the importance of context in network analysis and introduces innovative approaches like CubeSketch and Graph Zeppelin to enhance computation efficiency in handling complex graphs.

Dec 16, 2024 • 38min
Customizing a Graph Solution
Dave Bechberger, Principal Graph Architect at AWS and author of "Graph Databases in Action," shares his expertise on graph databases. He highlights their unique applications in fraud detection and the importance of analyzing complex networks. The discussion covers key algorithms like PageRank and community detection, offering insights into when organizations should transition from relational databases to graph solutions. Dave also addresses the challenges of skill shortages in the graph database field and shares practical strategies for successful implementation.

Dec 9, 2024 • 33min
Graph Transformations
Adam Machowczyk, a PhD student at the University of Leicester, specializes in graph rewriting and machine learning. He reveals how graph rewriting can enhance model adaptability, particularly in guiding machine learning for complex tasks. Topics include the transformation of graph structures for improved recommendations in social networks and its applications in chemistry and IoT analysis. Adam illustrates the shift from traditional data representation to dynamic graph systems, showcasing real-world implications and the future of scalable adaptive models.

6 snips
Nov 25, 2024 • 37min
Networks for AB Testing
Wentao Su, a data scientist at ByteDance, specializes in A/B testing for social media platforms. He dives into the challenges of A/B testing in dynamic networks, highlighting the spillover effects that can distort results. Wentao introduces innovative strategies like one-degree label propagation to optimize test accuracy. He also explores how user interconnectedness impacts both experimental design and user experience. Finally, he discusses the significance of robust data processing techniques to effectively manage large-scale experiments.

Nov 18, 2024 • 38min
Lessons from eGamer Networks
In this engaging discussion, Alex Bisberg, a PhD candidate at USC, dives into the fascinating world of network science and game analytics. He unveils how generosity spreads like a contagious virus within gaming communities and reveals the power of weak ties in fostering new connections. Bisberg explores the innovative use of candles as a unique in-game currency promoting collaborative play. Listeners will discover insights on social mechanics in games like 'Sky Children of the Light' and how these dynamics can enhance player engagement and retention.

7 snips
Nov 11, 2024 • 42min
Github Collaboration Network
Behnaz Moradi-Jamei, an assistant professor at James Madison University specializing in network data science, delves into the intricate web of GitHub contributors. She unveils her groundbreaking analysis of a sprawling network connecting 700,000 developers through shared contributions. The conversation touches on community detection algorithms, ethical considerations in network analysis, and innovative methodologies for enhancing collaboration insights. Behnaz emphasizes the importance of adapting algorithms to reflect real-world developer interactions, pushing the boundaries of open-source community understanding.

Nov 4, 2024 • 42min
Graphs and ML for Robotics
Join Abhishek Paudel, a PhD Student at George Mason University specializing in robotics and machine learning. He shares fascinating insights into how graph neural networks can classify rooms and enhance robotic navigation. Explore the evolution of machine learning in robotics, and the impact of deep learning on perception and motion control. Abhishek discusses the integration of natural language processing and innovative graph-based methods for decision-making, highlighting their role in improving spatial awareness and learning from mistakes.