4 papers
Approximate Computation via Le Cam Simulability
Deniz Akdemir
We propose a decision-theoretic framework for computational complexity, complementary to classical theory: moving from syntactic exactness (Turing / Shannon) to semantic simulabili…
TabMixNN: A Unified Deep Learning Framework for Structural Mixed Effects Modeling on Tabular Data
Deniz Akdemir
We present TabMixNN, a flexible PyTorch-based deep learning framework that synthesizes classical mixed-effects modeling with modern neural network architectures for tabular data an…
Le Cam Distortion: A Decision-Theoretic Framework for Robust Transfer Learning
Deniz Akdemir
Distribution shift is the defining challenge of real-world machine learning. The dominant paradigm--Unsupervised Domain Adaptation (UDA)--enforces feature invariance, aligning sour…
Likelihood-Preserving Embeddings for Statistical Inference
Deniz Akdemir
Modern machine learning embeddings provide powerful compression of high-dimensional data, yet they typically destroy the geometric structure required for classical likelihood-based…