Daniil Dorin

Decoding Visual Information from Neural Signals: Image Reconstruction Based on Joint fMRI and EEG Analysis

Daniil Dorin, Nikita Kiselev, Andrey Grabovoy

A multimodal architecture that jointly processes fMRI and EEG signals to reconstruct visual stimuli, with contrastive alignment to CLIP and a two-stage diffusion generation pipeline.

Evidential Image Matching: Predicting Transformation Sequences to Derive One Image from Another

Daniil Dorin, Kseniia Varlamova, Andrey Grabovoy

Plagiarism detection reformulated as predicting the sequence of transformations that derives one image from another, with an encoder–decoder model and the Canonical Jaccard Index.

Pairwise Image Matching for Plagiarism Detection

Daniil Dorin, Kseniia Varlamova, Andrey Grabovoy

A siamese network with a weight-shared encoder, symmetric fusion and a similarity head, trained with plagiarism-mimicking augmentations to minimise false positives in pairwise image plagiarism detection.

Enhancing fMRI Data Decoding with Spatiotemporal Characteristics in Limited Dataset

Daniil Dorin, Andrey Grabovoy, Vadim Strijov

An fMRI decoding methodology for small datasets that combines subject-specific brain activity masks with an encoder based on Riemannian geometry.

Forecasting fMRI Images From Video Sequences: Linear Model Analysis

Daniil Dorin*, Nikita Kiselev*, Andrey Grabovoy, Vadim Strijov

A method for approximating fMRI readings from the video sequence a person watches, based on a linear model for each voxel and a time-invariant hemodynamic response.

Decentralized Optimization with Coupled Constraints

Demyan Yarmoshik, Alexander Rogozin, Nikita Kiselev, Daniil Dorin, Alexander Gasnikov, Dmitry Kovalev

Lower complexity bounds for decentralized optimization with affine coupled constraints, and the first linearly convergent first-order decentralized algorithm that achieves them.

Just Relax It: discrete variable relaxation

PyTorch library of relaxation methods for discrete random variables in neural networks. Gumbel-Softmax Top-K, Hard Concrete, Straight-Through Bernoulli, Invertible Gaussian, REBAR and more, behind a Pyro-style distribution API.

HippoTrainer: gradient-based hyperparameter optimization

A PyTorch library that tunes hyperparameters by differentiating through the training loop. Implements T1-T2, Neumann-series implicit differentiation, HOAG and DrMAD behind one trainer interface.

Garage: generative augmentation framework

Replaces objects in images with newly generated ones. Grounded-SAM finds and masks the object, an Augmenter model proposes replacements and prompts, PowerPaint inpaints the result.

Epidemic spread models: COVID-19 as stochastic chemical kinetics

Various approaches to modeling the spread of epidemics, in particular COVID-19, through differential equations and Markov processes. Deterministic and stochastic SIR models fitted to real multi-wave data.