Accepted at ECAI'24 Workshop on Agents and Robots for Reliable Engineered Autonomy (AREA)
This paper introduces Attentive A* incorporating attention mechanism into data-driven path-planning algorithms. This algorithm makes use of SCSE mechanism along with VGG-19 in the encoder part of the algorithm to generate a guidance map. Download/View here
Preprint
This paper introduces the IndieFake Dataset (IFD), a comprehensive audio deepfake dataset featuring 27.12 hours of bonafide and deepfake audio from 50 Indian English speakers. Addressing the lack of South-Asian representation in existing datasets, IFD offers balanced data distribution and includes speaker-level characterization to enhance deepfake detection in diverse linguistic and cultural contexts. For details, visit the project page.
Preprint
This research focused on examining the origins of hate speech and the underlying reasons behind it. Tweets related to various controversial events were analyzed, with each event’s context carefully studied to ensure accurate interpretation. Each tweet was individually evaluated to determine whether it incited hate. Additionally, the study explored key differences between hate-instigating content and conventional hate speech. Download/View here