Sasanka Adikari’s Post

Awesome, finally published in the Journal of the American Statistical Association Theory and Methods! In this work, I address key computational challenges in Bayesian inference for graphical (decomposable) graphs. The introduced novel method uses latent dimensional expansion of Junction trees to come up with a parallel non-reversible MCMC sampling, yielding substantial efficiency gains (up to 10X), with added accuracy. Now you can compute partial correlations (conditional dependencies) for up to 1000s of variables in a matter of seconds ("under some assumptions"). Checkout Benchpress (https://lnkd.in/dQxURs8b) for comparative analysis. Article link: https://lnkd.in/d3K4hy_x #BayesianInference #GraphicalModels #MCMC #Statistics #BayesianStructureLearning #distributedLearning #BayesianNetworks

🔍 Just published! Delve into the world of graph labeling with this new research paper on "Antimagic Labeling of Graphs Using Prime Numbers." Explore the unique approach of assigning distinct prime values to edges, ensuring the cumulative sum of edge labels at each vertex remains unique. This fascinating study includes a conjecture on antimagic labeling of any graphs and proves two compelling theories. Read the full paper here: https://bit.ly/4azIvhR

I am delighted to share this thought-provoking blog post, "P not equal to NP." The article delves into the complex world of computational complexity, addressing the fundamental question of whether all NP problems are also P problems. By analyzing a specific decision problem, the post demonstrates that it falls within the class NP but not P, providing compelling evidence that not all NP problems are P. For a deeper understanding of this fascinating topic, click here: https://bit.ly/49aTLR1.

The paper identifies a phenomenon called "flips," where answers change from correct to incorrect and vice versa between the baseline and compressed models, even when overall accuracy remains similar. The study suggests that using accuracy alone is insufficient to evaluate compression techniques. The authors propose additional evaluation metrics, such as KL-Divergence and the percentage of flips, to better capture the behavioral differences between baseline and compressed models. Paper Link: https://lnkd.in/ezAaMygW

My new article, "Semi-Markov control models for systems of large populations of interacting objects with possible unbounded costs: a mean field approach", is now available online in the Annals of Operations Research Journal. So happy to start the week with this great news!!! https://lnkd.in/gvmf7twj

📢 Check out this paper that was published in our journal! 👉 The Influence of Herd Mentality on Rating Bias and Popularity Bias: A Bi-Process Debiasing Recommendation Model Based on Matrix Factorization 🔗 https://lnkd.in/d_qMwZpG

From the Blog:  In this post, I want to answer the question: Is it necessary to replicate the results of a published paper before publishing another paper on the same topic? To start with an answer, we first need to dive...

I’m excited to share my first published article https://lnkd.in/dwi9RhxX

I am pleased to share a comprehensive survey paper on the empirical and experimental perspectives of big data in recommendation systems. This paper analyzes current algorithms and proposes a novel, hierarchical taxonomy to thoroughly categorize them. The survey covers a wide range of algorithms and techniques, further evaluating them through empirical and experimental assessments. It also sheds light on the future prospects of big data techniques in recommendation systems. If you're interested in delving into this insightful analysis, you can access the full paper at https://bit.ly/3Sy6KWl.

Excited to share that my paper, 'Secret Key Extraction Using Keyloggers,' is now published in IJISRT! Check it out here: https://lnkd.in/dxFmk7eZ