6 papers
Think Shallow, Solve Deep: Controlling Recurrent Dynamics for Reliable Test-Time Depth
Ivan Viakhirev, Kirill Borodin, Amirah Almutairi +3
Recurrent-depth reasoners aim to solve harder problems by iterating their update longer at test time, but additional iterations can improve, preserve, or degrade an answer. We show…
The Null Token Knows: Reducing Message-Free Hallucination in ASR and NMT
Kirill Borodin, Vasiliy Kudryavtsev, Ivan Viakhirev +1
Modern encoder-decoder systems can produce fluent text even when their input contains no recoverable message. We study this failure in ASR and NMT through the models' reserved null…
Structural properties of the implicit function defined by an integral self-consistency equation
Ivan Viakhirev
We study the integral equation with , where is a probability density on vanishing polynomially at . Setting $\mathcal{I}^+…
From Dispersion to Attraction: Spectral Dynamics of Hallucination Across Whisper Model Scales
Ivan Viakhirev, Kirill Borodin, Grach Mkrtchian
Hallucinations in large ASR models present a critical safety risk. In this work, we propose the \textit{Spectral Sensitivity Theorem}, which predicts a phase transition in deep net…
Interpreting Multi-Branch Anti-Spoofing Architectures: Correlating Internal Strategy with Empirical Performance
Ivan Viakhirev, Kirill Borodin, Mikhail Gorodnichev +1
Multi-branch deep neural networks like AASIST3 achieve state-of-the-art comparable performance in audio anti-spoofing, yet their internal decision dynamics remain opaque compared t…
Towards Scalable AASIST: Refining Graph Attention for Speech Deepfake Detection
Ivan Viakhirev, Daniil Sirota, Aleksandr Smirnov +1
Advances in voice conversion and text-to-speech synthesis have made automatic speaker verification (ASV) systems more susceptible to spoofing attacks. This work explores modest ref…