4 papers
Leveraging Vision-Language Pre-training for Human Activity Recognition in Still Images
Cristina Mahanta, Gagan Bhatia
Recognising human activity in a single photo enables indexing, safety and assistive applications, yet lacks motion cues. Using 285 MSCOCO images labelled as walking, running, sitti…
Date Fragments: A Hidden Bottleneck of Tokenization for Temporal Reasoning
Gagan Bhatia, Maxime Peyrard, Wei Zhao
Modern BPE tokenizers often split calendar dates into meaningless fragments, e.g., 20250312 202, 503, 12, inflating token counts and obscuring the inherent structure…
DateLogicQA: Benchmarking Temporal Biases in Large Language Models
Gagan Bhatia, MingZe Tang, Cristina Mahanta +1
This paper introduces DateLogicQA, a benchmark with 190 questions covering diverse date formats, temporal contexts, and reasoning types. We propose the Semantic Integrity Metric to…
Swan and ArabicMTEB: Dialect-Aware, Arabic-Centric, Cross-Lingual, and Cross-Cultural Embedding Models and Benchmarks
Gagan Bhatia, El Moatez Billah Nagoudi, Abdellah El Mekki +2
We introduce {\bf Swan}, a family of embedding models centred around the Arabic language, addressing both small-scale and large-scale use cases. Swan includes two variants: Swan-Sm…