A Power Talk on the topic “Discovering Patterns in Data: How AI Understands Us” was organized on 11th October 2025 from 9.00 a.m. to 5.30 p.m. The session was conducted by Mr. Dharma Teja, Research Scholar from the National Institute of Technology, Tiruchirappalli (NITT), and was introduced by Mr. S. Manikandan, Assistant Professor, GSB. The event aimed to provide deep insights into how Artificial Intelligence (AI) techniques help in discovering patterns hidden within complex data.
Mr. Teja. started with an explanation of Multivariate Regression Analysis, highlighting how multiple variables can be used to predict outcomes and how regression models are applied in business forecasting, finance, and research. He emphasised the importance of interpreting coefficients and evaluating model performance through measures like R² and adjusted R².
He then discussed various classification techniques, including Decision Trees, k-Nearest Neighbour (k-NN), Logistic Regression, and Discriminant Analysis. In Decision Trees, he explained how data is split based on specific attributes using criteria like information gain and Gini impurity. The k-NN algorithm was introduced as a simple yet powerful non-parametric approach that classifies data based on proximity to labelled instances. Through Logistic Regression, he explained how probabilities can be modelled for binary outcomes using the sigmoid function. Discriminant Analysis, both linear and quadratic, was discussed as a statistical technique for separating classes based on covariance structures. In the afternoon session, Mr. Teja moved on to Clustering techniques, an essential part of unsupervised learning. He explained how clustering helps in identifying groups or segments within data without predefined labels. Algorithms such as K-Means, Hierarchical Clustering, and DBSCAN were discussed in detail, with real-world examples like customer segmentation and anomaly detection. Participants gained a better understanding of how clustering is used to uncover meaningful patterns from raw data.
The final session focused on Neural Networks, where Mr. Teja elaborated on how these models simulate the working of the human brain through interconnected nodes and layers. He discussed the structure of neural networks, including input, hidden, and output layers, and explained how learning occurs through backpropagation and gradient descent optimization. Examples of deep learning applications, such as image and speech recognition, were shared to demonstrate the power of neural networks in real-world AI systems. The entire session was highly interactive, with participants actively engaging in discussions and asking questions about practical applications of the concepts. Mr Teja emphasised the importance of combining theoretical understanding with hands-on data analysis using modern AI tools.
The Power Talk on “Discovering Patterns in Data: How AI Understands Us” was a highly informative and inspiring event. The participants expressed their appreciation for the session, which enhanced their understanding of data analytics, machine learning, and the role of AI in uncovering meaningful patterns from data.