Ayan Das
Biography
Results-driven engineering professional with a strong academic foundation and hands-on expertise in project site management, analog/digital circuit design, and electrical simulations. Recently completed M.Tech in Electronics (2022?2024), bringing proven experience in coordinating field operations, ensuring compliance, and executing technical plans efficiently. Skilled in EDA tools, Python-based simulations, and technical reporting, with a consistent focus on quality, safety, and optimized project timelines.
Research Interest
AI-Based Emotional Archaeology: A Multilingual Transformer Framework for Reconstructing Historical Sentiment and Societal Psyche from Ancient Texts
Abstract
History often records battles, empires, and economies?but rarely the emotional undercurrents that defined the collective consciousness of past civilizations. This work proposes an innovative paradigm: AI-Based Emotional Archaeology?the reconstruction of historical emotional landscapes using multilingual, context-aware transformer models. Our system, SentivoX, is designed to extract emotional patterns, sentiment arcs, and societal moods from ancient texts spanning multiple languages, epochs, and cultural domains. The novelty of this research lies in integrating deep learning with historical context modeling, enabling the interpretation of subjective emotional states encoded within classical literature, religious scriptures, political treatises, and oral traditions. SentivoX utilizes a hybrid training pipeline: - Stage I: Pretraining on a multilingual corpus combining contemporary emotional datasets with translated ancient texts to align emotional semantics across time. - Stage II: Fine-tuning with cultural ontologies, genre-specific classifiers (e.g., epic, legal, ritual), and temporal attention mechanisms to mitigate modern bias. - Stage III: Application of the model to real-world classical sources, generating emotional timelines and confidence-rated interpretations. We demonstrate the effectiveness of SentivoX through four cross-civilizational case studies: 1. Sanskrit Epics (Mahabharata, Bhagavad Gita): Emotional polarity transitions during moral conflicts. 2. Classical Greek Philosophy: Extraction of anxiety and existentialism in pre-Socratic thought. 3. Latin Political Speeches: Tracing rhetorical emotion in Roman Senate records. 4. Mayan Codices and Ritual Texts: Identification of awe, reverence, and fear in ceremonial invocations. To visualize findings, we introduce the Temporal Emotion Map (TEM)?a novel tool that represents emotion valence and intensity along historical timelines, highlighting cultural inflection points such as wars, religious reformations, or plagues. Our results show that AI models, when carefully adapted with cultural knowledge and temporal logic, can recover emotionally nuanced narratives often overlooked by conventional historiography. The proposed framework opens a new interdisciplinary field intersecting AI, anthropology, digital humanities, and history, with applications in emotional heritage preservation, digital archiving, and cultural analytics.
Keywords: Emotional archaeology, NLP, ancient texts, sentiment analysis, transformer models, digital humanities, cross-cultural AI, historical emotion modeling
