5 papers
Why Clipping Matters in AdaGrad? Toward a High-Probability Theory under Generalized Smoothness
Alokendu Mazumder, Ayaan Mohd, Harshit Rawat +3
We analyze the original same-step coordinate-wise AdaGrad under generalized smoothness and heavy-tailed noise with bounded variance. In this setting, local curvature may grow sub-q…
On Hamming-Lipschitz Type Stability of the Subdominant (Minmax) Ultrametric: Theory and Simple Proofs
Alokendu Mazumder, Arnab Roy, Punit Rathore
The subdominant (minmax) ultrametric is a canonical tree-structured summary of a dissimilarity matrix, arising equivalently as the ultrametric induced by single-linkage clustering.…
Learning Low-Rank Latent Spaces with Simple Deterministic Autoencoder: Theoretical and Empirical Insights
Alokendu Mazumder, Tirthajit Baruah, Bhartendu Kumar +3
The autoencoder is an unsupervised learning paradigm that aims to create a compact latent representation of data by minimizing the reconstruction loss. However, it tends to overloo…
A Theoretical and Empirical Study on the Convergence of Adam with an "Exact" Constant Step Size in Non-Convex Settings
Alokendu Mazumder, Rishabh Sabharwal, Manan Tayal +2
In neural network training, RMSProp and Adam remain widely favoured optimisation algorithms. One of the keys to their performance lies in selecting the correct step size, which can…
DeepVAT: A Self-Supervised Technique for Cluster Assessment in Image Datasets
Alokendu Mazumder, Tirthajit Baruah, Akash Kumar Singh +3
Estimating the number of clusters and cluster structures in unlabeled, complex, and high-dimensional datasets (like images) is challenging for traditional clustering algorithms. In…