Tanat Tonguthaisri, CISSP®’s Post

🌟 Excited to share our latest blog post on survival modeling using deep learning and statistical methods! In this comparative analysis, we delved into various survival analysis techniques, from traditional statistical models to cutting-edge machine learning algorithms, assessing their performance in predicting mortality after hospital admission. The study included an in-depth comparison of methods such as Cox proportional hazards, Random Survival Forests, Gradient Boosting machine learning, and DeepSurv, among others. Our findings shed light on the effectiveness of deep learning in achieving comparable performance, with DeepSurv emerging as the top performer. Discover more insights and the full analysis here: https://bit.ly/3IARCTp #SurvivalModeling #DeepLearning #HealthcareAnalytics

Our latest article breaks down 10 key terms you need to know to navigate the world of data with confidence! 📊 Read: https://lnkd.in/gsMx-GG2 From cleaning messy datasets to unraveling the mysteries of Deep Learning, we've got you covered!

🌟 Excited to share our latest blog post on "Back to the Basics on Predicting Transfer Performance"! In the dynamic world of deep learning, choosing the optimal pre-trained models presents a formidable challenge. Our research delves into the proliferation of transferability scorers and the need for robust benchmark guidelines to evaluate them effectively. We introduce a compelling technique to combine multiple scorers, consistently enhancing their predictive results. Our thorough evaluation of 13 scorers across 11 diverse datasets illustrates the potential of combining information sources for reliable transferability prediction. Dive into the details at: https://bit.ly/4c5U699 #TransferLearning #DeepLearning #DataScience

Just having a little fun this mid-week with our Deep Learning Analytics team. Did you know? 🤔💭 We develop machine learning technologies that support smarter forecasting, threat detection, complicated problem solving and aid in decision-making. These data driven technologies give warfighters on the front lines access to the power of Artificial Intelligence (AI). Find out more about what our Deep Learning Analytics team is up to at: https://lnkd.in/gy3rhH7R

Hello everyone! On Day 27 of our #100DaysOfAI challenge, we will explore some of the most influential and popular architectures in Deep Learning. These architectures have paved the way for advancements in the field, with applications ranging from image classification to real-time object detection. Let's discover them! You can find more details and resources in Spanish on my GitHub repository: https://lnkd.in/e26tPcU8

Check out this short video to see some highlights from the session, including the students’ reactions and key takeaways on AI, ML, and Data Science! 📹

Finally, my article "Generalized naming game and Bayesian naming game as dynamical systems" is published. Among the results, it shows that there exists a nongeneric bifurcation in the generalised naming game, a classical model of semiotic dynamics, and how such a bifurcation modifies the noise of the stochastic sample paths involved. In this regard, it is found that there are two possible stochastic processes: the Brownian motion and the Ornstein-Uhlenbeck process. Finally, it shows that the logistic function, widely used in Deep Learning, can be employed for modeling the word learning processes according to a Bayesian concept learning framework proposed by Josh Tenenbaum. As a result, such an approach introduces a symmetry breaking mechanism, otherwise absent in the Bayesian dynamical system. Please, have a look at https://lnkd.in/dGaDV8Ac

A few weeks ago Ariel Kapusta and I published "On the Out-of-Distribution Coverage of Combining Split Conformal Prediction and Bayesian Deep Learning" at Transactions of Machine Learning Research (https://jmlr.org/tmlr/). I think it will be informative for anyone using (or planning to use) conformal prediction with their deep learning models. Check it out here: https://lnkd.in/e7XsMExV

It all started with N-BEATS in 2020, which was followed by NHITS in 2022. In 2023, PatchTST and TSMixer were proposed, and they still rank among the top forecasting models. More recently, discovered the iTransformer, which further pushed the performance of deep learning forecasting models. Now, we introduce the Series-cOre Fused Time Series forecaster or SOFTS.

I've been exploring innovative approaches to dimension reduction, moving beyond traditional techniques like PCA to delve into methods such as NMF, t-SNE, and UMAP. Recently, I discovered something even more exciting: Deep-TDA (Deep Targeted Discriminant Analysis). This novel approach combines the insights of Topological Data Analysis (TDA) with the power of deep learning, offering a unique lens through which to examine complex datasets. By focusing on the 'shape' of data, Deep-TDA unveils patterns that were previously hidden. I recently came across an insightful article about Deep-TDA that has greatly inspired me. It's definitely worth checking out: https://lnkd.in/gbWAk3ky