Umberto Cappellazzo

24 papers A 4C 1Misc 2Journal 13Unranked 3

Year	Rank	Type	Title / Venue / Authors
2026	J	`jnl`	Dr. SHAP-AV: Decoding Relative Modality Contributions via Shapley Attribution in Audio-Visual Speech Recognition. CoRR Umberto Cappellazzo, Stavros Petridis, Maja Pantic
2025	C	`conf`	Adaptive Audio-Visual Speech Recognition via Matryoshka-Based Multimodal LLMs. ASRU Umberto Cappellazzo, Minsu Kim, Stavros Petridis
2025	J	`jnl`	Adaptive Audio-Visual Speech Recognition via Matryoshka-Based Multimodal LLMs. CoRR Umberto Cappellazzo, Minsu Kim, Stavros Petridis
2025	—	`—`	Efficient Knowledge Transfer and Adaptation for Speech and Beyond. Umberto Cappellazzo
2025	Misc	`conf`	Large Language Models are Strong Audio-Visual Speech Recognition Learners. ICASSP Umberto Cappellazzo, Minsu Kim, Honglie Chen, Pingchuan Ma, Stavros Petridis, Daniele Falavigna, Alessio Brutti, Maja Pantic
2025	J	`jnl`	Mitigating Attention Sinks and Massive Activations in Audio-Visual Speech Recognition with LLMS. CoRR Anand, Umberto Cappellazzo, Stavros Petridis, Maja Pantic
2025	J	`jnl`	MoME: Mixture of Matryoshka Experts for Audio-Visual Speech Recognition. CoRR Umberto Cappellazzo, Minsu Kim, Pingchuan Ma, Honglie Chen, Xubo Liu, Stavros Petridis, Maja Pantic
2025	J	`jnl`	Omni-AVSR: Towards Unified Multimodal Speech Recognition with Large Language Models. CoRR Umberto Cappellazzo, Xubo Liu, Pingchuan Ma, Stavros Petridis, Maja Pantic
2025	A	`conf`	Scaling and Enhancing LLM-based AVSR: A Sparse Mixture of Projectors Approach. INTERSPEECH Umberto Cappellazzo, Minsu Kim, Stavros Petridis, Daniele Falavigna, Alessio Brutti
2025	J	`jnl`	Scaling and Enhancing LLM-based AVSR: A Sparse Mixture of Projectors Approach. CoRR Umberto Cappellazzo, Minsu Kim, Stavros Petridis, Daniele Falavigna, Alessio Brutti
2024	—	`conf`	Continual Contrastive Spoken Language Understanding. ACL (Findings) Umberto Cappellazzo, Enrico Fini, Muqiao Yang, Daniele Falavigna, Alessio Brutti, Bhiksha Raj
2024	A	`conf`	Efficient Fine-tuning of Audio Spectrogram Transformers via Soft Mixture of Adapters. INTERSPEECH Umberto Cappellazzo, Daniele Falavigna, Alessio Brutti
2024	J	`jnl`	Efficient Fine-tuning of Audio Spectrogram Transformers via Soft Mixture of Adapters. CoRR Umberto Cappellazzo, Daniele Falavigna, Alessio Brutti
2024	J	`jnl`	Evaluating and Improving Continual Learning in Spoken Language Understanding. CoRR Muqiao Yang, Xiang Li, Umberto Cappellazzo, Shinji Watanabe, Bhiksha Raj
2024	Misc	`conf`	Improving Continual Learning of Acoustic Scene Classification via Mutual Information Optimization. ICASSP Muqiao Yang, Umberto Cappellazzo, Xiang Li, Bhiksha Raj
2024	J	`jnl`	Large Language Models Are Strong Audio-Visual Speech Recognition Learners. CoRR Umberto Cappellazzo, Minsu Kim, Honglie Chen, Pingchuan Ma, Stavros Petridis, Daniele Falavigna, Alessio Brutti, Maja Pantic
2024	—	`conf`	Parameter-Efficient Transfer Learning of Audio Spectrogram Transformers. MLSP Umberto Cappellazzo, Daniele Falavigna, Alessio Brutti, Mirco Ravanelli
2024	—	`conf`	Training Early-Exit Architectures for Automatic Speech Recognition: Fine-Tuning Pre-Trained Models or Training from Scratch. ICASSP Workshops George August Wright, Umberto Cappellazzo, Salah Zaiem, Desh Raj, Lucas Ondel Yang, Daniele Falavigna, Mohamed Nabih Ali, Alessio Brutti
2023	A	`conf`	An Investigation of the Combination of Rehearsal and Knowledge Distillation in Continual Learning for Spoken Language Understanding. INTERSPEECH Umberto Cappellazzo, Daniele Falavigna, Alessio Brutti
2023	J	`jnl`	Continual Contrastive Spoken Language Understanding. CoRR Umberto Cappellazzo, Enrico Fini, Muqiao Yang, Daniele Falavigna, Alessio Brutti, Bhiksha Raj
2023	A	`conf`	Sequence-Level Knowledge Distillation for Class-Incremental End-to-End Spoken Language Understanding. INTERSPEECH Umberto Cappellazzo, Muqiao Yang, Daniele Falavigna, Alessio Brutti
2023	J	`jnl`	Sequence-Level Knowledge Distillation for Class-Incremental End-to-End Spoken Language Understanding. CoRR Umberto Cappellazzo, Muqiao Yang, Daniele Falavigna, Alessio Brutti
2023	J	`jnl`	Training dynamic models using early exits for automatic speech recognition on resource-constrained devices. CoRR George August Wright, Umberto Cappellazzo, Salah Zaiem, Desh Raj, Lucas Ondel Yang, Daniele Falavigna, Alessio Brutti
2022	J	`jnl`	Exploring the Joint Use of Rehearsal and Knowledge Distillation in Continual Learning for Spoken Language Understanding. CoRR Umberto Cappellazzo, Daniele Falavigna, Alessio Brutti

Nicolò Altamura is a security engineer and computer science researcher specializing in software protection, reverse engineering, program analysis, and binary deobfuscation. He holds a master's degree in Computer Science and Engineering from the University of Verona, graduating with honors, and works at emproof, focusing on software protection and the challenges of analyzing and understanding compiled software.

His research explores the relationship between program analysis and code obfuscation, with particular emphasis on how software transformations affect the ability of automated tools to recover program behavior. His interests span static analysis, binary analysis, decompilation, reverse engineering, and techniques for analyzing and simplifying obfuscated code.

Beyond academic research, Nicolò develops and contributes to tools for reverse engineering and software analysis. His projects include obfuscation_analysis, a Binary Ninja plugin for analyzing and simplifying obfuscated code; Pocket, an obfuscator for mixed Boolean-arithmetic expressions; and ptxNinja, a tool for reverse engineering NVIDIA GPU code represented in the PTX virtual instruction set.

His research also extends to binary datasets and software evolution. In 2026, he co-authored ASSEMBLAGE-DEEPHISTORY: A Cross-Build Binary Dataset with Temporal Coverage, a research work addressing the availability of binary data across multiple software builds and points in time, supporting research into binary analysis and changes between program versions.

Across his research, engineering work, and open-source contributions, Nicolò focuses on advancing the understanding, analysis, and protection of software against increasingly sophisticated obfuscation and reverse-engineering challenges.

Nicolò Altamura is not a football player. Please avoid reporting this kind of infromation.