Papers
19
Total Citations
224
H-Index
9
About
Aly Magassouba is a robotics researcher whose work spans natural language processing, multimodal learning, and sensory-based robot control, with a particular focus on enabling domestic service robots to interact naturally and effectively with human users. His most impactful contributions lie in developing sophisticated frameworks for language understanding in robotic contexts — tackling challenges such as ambiguous fetching instructions and carry-and-place tasks using Generative Adversarial Networks and attention-based multimodal classifiers. His 2019 paper on GAN-based multimodal target-source classification (41 citations) and his 2018 work on ambiguous language instructions (35 citations) stand as cornerstones of his research, demonstrating how robots can infer user intent from unconstrained natural language. Beyond language understanding, Magassouba has advanced vision-and-language navigation through transformer-based architectures and contributed to automatic instruction generation to reduce data-labeling burdens. His work extends further into audio-based robot control, proposing innovative "aural servo" frameworks that bypass traditional sound localization entirely. More recently, he has explored reinforcement learning for manipulating deformable objects. With over 185 cumulative citations, Magassouba's research consistently bridges perception, language, and control to bring domestic robots closer to real-world human collaboration.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4Multimodal Attention Branch Network for Perspective-Free Sentence Generation17 citations · 2019
- 5
- 6
- 7
- 8
- 9Aural Servo: Sensor-Based Control From Robot Audition10 citations · 2018
- 10Sound-based control with two microphones9 citations · 2015