Mansi Sharma

Timeline

1. 2020

   Joining DFKI in Saarbrücken

   She joins DFKI in 2020 as a researcher and PhD student in computer science. Before that, a Graduate Fellowship from Saarland University supports her research, and her degrees are a master's in computer science from Saarland University and a bachelor's from NIT Raipur, where she receives a departmental silver medal for academic excellence.

2. 2020–2021

   Decoding semantic categories from EEG

   She is among the coauthors of two studies on decoding semantic categories from EEG activity: one on object-based decision tasks at the 8th International Winter Conference on Brain-Computer Interface in 2020, and one on silent speech imagination tasks at the 9th in 2021.

3. 2022

   An EEG search-intent pipeline

   In 2022, she coauthors "Towards Improving EEG-based Intent Recognition in Visual Search Tasks" for the International Conference on Neural Information Processing (ICONIP) with Maurice Rekrut, Jan Alexandersson and Antonio Krüger. The end-to-end pipeline separates navigational intent, looking at a scene, from informational intent, searching for a target, and reaches 97.89% mean accuracy in the subject-specific case; a cross-subject version reaches 96.83%.

4. 2023

   A public EEG and gaze dataset

   In 2023, she coauthors "Implicit Search Intent Recognition using EEG and Eye Tracking: Novel Dataset and Cross-User Prediction" with Shuang Chen, Philipp Müller, Maurice Rekrut and Antonio Krüger for the 25th International Conference on Multimodal Interaction (ICMI). The paper offers the first publicly available EEG and eye-tracking dataset for navigational versus informational intent recognition, in which users decide how long they search, and its cross-user method reaches 84.5% accuracy in leave-one-user-out evaluations, close to the 85.5% of within-user prediction.

5. 2023

   Intent in robot-assisted surgery

   Her 2023 paper with Antonio Krüger at the IEEE International Conference on Systems, Man, and Cybernetics (SMC), "Unraveling the Connection: How Cognitive Workload Shapes Intent Recognition in Robot-Assisted Surgery", describes a project that builds an intelligent adaptive system to monitor cognitive workload from brain activity, heart rate, muscle activity and eye tracking. Her DFKI project on cognitive workload in intent recognition during robot-assisted surgery, SC_INCOW, has the Zeiss group as partner. She also visits NII Japan as an intern in 2023.

6. 2024

   Target and non-target fixations

   In 2024, she publishes with Camilo Martínez, Benedikt Emanuel Wirth, Antonio Krüger and Philipp Müller, in the proceedings of the 26th International Conference on Multimodal Interaction (ICMI), the first study of target versus non-target fixation classification during free visual search in realistic scenes. The 36-participant study with 140 scenes reaches 83.6% accuracy in cross-user evaluations from gaze and EEG features, where methods based on saccade-related potentials reach 56.9%.

7. October 2025

   Startle signals and Bitcoin forecasting

   In October 2025, she publishes "Distinguishing Startle from Surprise Events Based on Physiological Signals" at the IEEE MetroXRAINE conference in Ancona, with Alexandre Duchevet, Florian Daiber, Jean-Paul Imbert and Maurice Rekrut. Using machine learning and multimodal fusion, the study separates startle from surprise events with 85.7% mean accuracy. Her 2025 ICONIP paper with Enrico Sartor, Marc Cavazza and Helmut Prendinger applies Echo State Networks to Bitcoin time series prediction.

8. As of October 2026

   Doctoral researcher at DFKI

   She is a researcher and doctoral candidate in the Cognitive Assistants department at DFKI and Saarland University, and has been at DFKI since 2020. Her work on human intent recognition integrates EEG, eye tracking and gestures in visual search tasks and microsurgical interactions. She is among the authors of "Tools, Challenges and Opportunities to Address Intra- and Inter-User Variability in EEG-based BCIs" for the 34th European Signal Processing Conference (EUSIPCO 2026).