CV

ITALIAN INSTITUTE OF TECHNOLOGY (IIT) | Postdoctoral Researcher

Multimodal learning and spatial audio for immersive environments in the European XTREME project.

• Developing few-shot room impulse response generation from audio and depth, with counterfactual training for reverberation control (ongoing research).
• Built Eigenmike 64 processing pipelines for beamforming, binaural rendering, and acoustic mapping on 3D room point clouds.
• Implemented audio/video synchronization and 360-degree acoustic overlays for spatial-audio recordings.

UNIVERSITY OF GENOA | PhD Researcher

Research on multimodal human motion generation and domain generalization, including a research collaboration with King's College London.

• Designed ALaDiT, a multimodal diffusion transformer generating 4 s of 3D motion in <14 ms on a single RTX 3090 (SICAGE, ECCV 2026).
• Created and released TED4C-L: 106.45 hours from 764 speakers, aligning MMPose-extracted motion with audio and text.
• Analyzed cultural cues in interaction and affective data using classical and deep learning models (AIRO 2022; AHRI at ACII 2024).
• Learned cultural representations with Fishr and adversarial domain generalization, using speaker-disjoint training and evaluation.
• Benchmarked generative models with controlled ablations; released training and evaluation code.
• Developed semantic-similarity algorithms for symbolic and deictic gesture labeling (IEEE RO-MAN 2024).

AKKA TECHNOLOGIES | Software Engineer

• Developed Python backend software for aircraft panels and presented a working demo to stakeholders.

Ph.D. in Artificial Intelligence and Robotics

University of Genoa

• Research collaboration with King's College London; thesis defended in May 2025.

M.Sc. in Robotics Engineering

University of Genoa

• Thesis: Culture-Aware Co-Speech Gestures Using Generative Adversarial Networks.
• Developed speech-driven GAN gesture generation tested on the Pepper robot (AIRO 2021; International Journal of Social Robotics 2022).

B.Sc. in Biomedical Engineering

University of Genoa

SICAGE: Speaker-Independent Culture-Aware Gesture Generation using TED4C-L Dataset

European Conference on Computer Vision (ECCV), main conference, pp. 392-411

Ariel Gjaci, Antonio Sgorbissa, Vittorio Murino

Labeling Sentences with Symbolic and Deictic Gestures via Semantic Similarity

IEEE RO-MAN, pp. 477-484

Ariel Gjaci, Carmine Tommaso Recchiuto, Antonio Sgorbissa

Exploring Cultural Cues in Multimodal Interactions

AHRI Workshop at ACII, pp. 306-313

Ariel Gjaci, Viktor Schmuck, Antonio Sgorbissa, Oya Celiktutan

Culture Awareness in Intelligent Systems

AIRO Workshop

Ariel Gjaci, Luca Oneto, Carmine Tommaso Recchiuto, Antonio Sgorbissa

Towards Culture-Aware Co-Speech Gestures for Social Robots

International Journal of Social Robotics, 14(6), pp. 1493-1506

Ariel Gjaci, Carmine Tommaso Recchiuto, Antonio Sgorbissa

A GAN-based Approach for Generating Culture-Aware Co-Speech Gestures

AIRO Workshop

Ariel Gjaci, Carmine Tommaso Recchiuto, Antonio Sgorbissa

Teaching and Academic Service

University of Genoa and research community

Teaching assistant, workshop co-organizer, reviewer, and session chair.

• Teaching assistant: Foundations of Informatics (C++), 60 hours; Mobile Programming (Android/Kotlin), 30 hours.
• Co-organizer of the ECCV workshop Expressive Encounters.
• Reviewer for AAAI, ECCV, ACM Multimedia, IROS, ICMI, and the International Journal of Social Robotics.
• Session chair at IEEE RO-MAN 2024.