Rahul Raja

Rahul Raja


Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
23 August 2024 in Preprints.org
Verified
Preprint
Open access
PDF file
AVArpita Vats
Arpita Vats
RRRahul Raja
Rahul Raja
VJVinija Jain
Vinija Jain
ACAman Chadha
Aman Chadha
The Mixture of Experts (MoE) architecture has evolved as a powerful and versatile approach for improving the performance and efficiency of deep learning models. This survey aims to provide the fundamental principles of MoE in details, a paradigm that harnesses the collective power of multiple specialized "expert" models working in concert to tackle complex tasks.Our exploration is a thorough analysis of the core components that constitute the MoE framework. We begin by dissecting the routing mechanism, a crucial element responsible for dynamically assigning input data to the most appropriate expert models. This routing process is pivotal in ensuring that each expert's specialized knowledge is optimally utilized. The survey places significant emphasis on expert specialization, a key feature that sets MoE apart from traditional architectures. We examine various strategies for developing and training specialized experts, exploring how this specialization enables MoE models to effectively handle diverse and multifaceted problems.Load balancing, another critical aspect of MoE systems, receives thorough attention. We discuss techniques for efficiently distributing computational resources among experts, ensuring optimal model performance while managing hardware constraints. This section provides insights into the delicate balance between model capacity and computational efficiency. Furthermore, we provide an in-depth discussion of the expert models themselves, examining their architectural designs, training methodologies, and how they interact within the larger MoE framework. This includes an analysis of different types of experts, from simple neural networks to more complex, task-specific architectures. The survey also navigates through diverse research avenues within the MoE landscape, highlighting recent advancements and innovative applications across various domains of machine learning. We pay particular attention to the burgeoning use of MoE in two rapidly evolving fields: computer vision and large language model (LLM) scaling. By providing this comprehensive overview, our survey aims to offer researchers and practitioners a deep understanding of MoE's capabilities, current applications, and potential future directions in the ever-evolving landscape of deep learning.The Mixture of Experts (MoE) architecture has evolved as a powerful and versatile approach for improving the performance and efficiency of deep learning models. This survey aims to provide the fundamental principles of MoE in details, a paradigm that harnesses the collective power of multiple specialized "expert" models working in concert to tackle complex tasks.Our exploration is a thorough analysis of the core components that constitute the MoE framework. We begin by dissecting the routing mechanism, a crucial element responsible for dynamically assigning input data to the most appropriate expert models. This routing process is pivotal in ensuring that each expert's specialized knowledge is optimally utilized. The survey places significant emphasis on expert specialization, a key feature that sets MoE apart from traditional architectures. We examine various strategies for developing and training specialized experts, exploring how this specialization enables MoE models to effectively handle diverse and multifaceted problems.Load balancing, another critical aspect of MoE systems, receives thorough attention. We discuss techniques for efficiently distributing computational resources among experts, ensuring optimal model performance while managing hardware constraints. This section provides insights into the delicate balance between model capacity and computational efficiency. Furthermore, we provide an in-depth discussion of the expert models themselves, examining their architectural designs, training methodologies, and how they interact within the larger MoE framework. This includes an analysis of different types of experts, from simple neural networks to more complex, task-specific architectures. The survey also navigates through diverse research avenues within the MoE landscape, highlighting recent advancements and innovative applications across various domains of machine learning. We pay particular attention to the burgeoning use of MoE in two rapidly evolving fields: computer vision and large language model (LLM) scaling. By providing this comprehensive overview, our survey aims to offer researchers and practitioners a deep understanding of MoE's capabilities, current applications, and potential future directions in the ever-evolving landscape of deep learning.

2 October 2025 in Preprints.org
Verified
Preprint
Open access
PDF file
VPVishnu Pendyala
Vishnu Pendyala
RRRahul Raja
Rahul Raja
AVArpita Vats
Arpita Vats
RPRaghu Para
Raghu Para
DKDurga Krishnamoorthy
Durga Krishnamoorthy
SLShruti Lohani
Shruti Lohani
This paper introduces the October 2025 issue of The Cognitive Nexus Magazine, which brings together six diverse contributions at the intersection of artificial intelligence theory, practice, and application. The featured works span methodological innovation, applied frameworks, and infrastructure design. At the methodological frontier, the issue highlights advances in causal neuro-symbolic reinforcement learning that integrate structural causal models with temporal logic for robust and interpretable decision-making under uncertainty. Applied perspectives are represented through a real-time AI-driven retail eCommerce and point-of-sale optimization framework, demonstrating measurable improvements in demand forecasting, operational efficiency, and customer engagement. From an infrastructure standpoint, the issue explores architectures for multi-tenant AI platforms that balance efficiency with performance isolation. Complementary articles expand the scope further, addressing emerging domains and cross-disciplinary intersections. Collectively, the contributions underscore three themes shaping the AI research and innovation landscape today: robustness and generalization in learning systems, scalability and efficiency in deployment, and ethical governance in applied settings. By synthesizing these threads, this issue provides both researchers and practitioners with insights into current challenges and future directions in the design of trustworthy and scalable AI systems

23 September 2025 in Preprints.org
Verified
Preprint
Open access
PDF file
RRRahul Raja
Rahul Raja
AVArpita Vats
Arpita Vats
Sudipta RoySudipta Roy
Sudipta Roy
Recent advancements in Large Language Models (LLMs) have enabled a paradigm shift in recommender systems, moving from modular pipelines toward instruction driven recommendation via prompt engineering. While existing literature has explored LLMs for tasks such as candidate generation, conversational recommendation, and re ranking, there remains a lack of systematic understanding of how prompts can be designed to optimize for distinct ranking objectives beyond relevance such as diversity, novelty, serendipity, and fairness. In this survey, we present a comprehensive review of prompt engineering techniques tailored for LLM based recommender systems, with a focus on ranking optimization under multi objective settings. We first introduce a taxonomy of prompt design strategies ranging from zero shot instruction templates to few shot exemplars and chain of thought prompting across different stages of recommendation (generation, ranking, re ranking). We then examine how these prompts can be aligned with specific ranking goals, and evaluate the tradeoffs between static prompting, prompt tuning, and fine tuning approaches. We review recent empirical studies and identify open challenges in prompt generalization, robustness, prompt evaluation protocols, and the absence of standardized benchmarks for multi objective recommendation tasks. Our survey concludes with actionable research directions and proposes a unified framework for evaluating prompt effectiveness across ranking objectives in LLM based recommender systems.

3 November 2025 in Preprints.org
Verified
Preprint
Open access
PDF file
AVArpita Vats
Arpita Vats
RRRahul Raja
Rahul Raja
AKAshish Kattamuri
Ashish Kattamuri
ABAbhinav Bohra
Abhinav Bohra
Quantum Natural Language Processing (QNLP) is an emerging interdisciplinary field at the intersection of quantum computing, natural language understanding, and formal linguistic theory. As advances in quantum hardware and algorithms accelerate, QNLP promises new paradigms for representation learning, semantic modeling, and efficient computation. However, existing literature remains fragmented, with no unified synthesis across modeling, encoding, and evaluation dimensions.In this work, we present the first systematic and taxonomy driven survey of QNLP that holistically organizes research spanning three core dimensions: computational models, encoding paradigms, and evaluation frameworks. First, we analyze foundational approaches that map linguistic structures into quantum formalism, including categorical compositional models, variational quantum circuits, and hybrid quantum classical architectures. Second, we introduce a unified taxonomy of encoding strategies, ranging from quantum tokenization and state preparation to embedding based encodings, highlighting tradeoffs in scalability, noise resilience, and expressiveness. Third, we provide the first comparative synthesis of evaluation methodologies, benchmark datasets, and performance metrics, while identifying reproducibility and standardization gaps.We further contrast quantum inspired NLP methods with fully quantum implemented systems, offering insights into resource efficiency, hardware feasibility, and real world applicability. Finally, we outline open challenges such as integration with LLMs and unified benchmark design, and propose a research agenda for advancing QNLP as a scalable and reliable discipline.

29 September 2025 in Preprints.org
Verified
Preprint
Open access
PDF file
AKAshish Kattamuri
Ashish Kattamuri
HFHarshwardhan Fartale
Harshwardhan Fartale
AVArpita Vats
Arpita Vats
RRRahul Raja
Rahul Raja
Data contamination poses a significant challenge to reliable LLM evaluation, where models may achieve high performance by memorizing training data rather than demonstrating genuine reasoning capabilities. We introduce RADAR (Recall vs. Reasoning Detection through Activation Representation), a novel framework that leverages mechanistic interpretability to detect contamination by distinguishing recall-based from reasoning-based model responses. RADAR extracts 37 features spanning surfacelevel confidence trajectories and deep mechanistic properties including attention specialization, circuit dynamics, and activation flow patterns. Using an ensemble of classifiers trained on these features, RADAR achieves 93% accuracy on a diverse evaluation set, with perfect performance on clear cases and 76.7% accuracy on challenging ambiguous examples. This work demonstrates the potential of mechanistic interpretability for advancing LLM evaluation beyond traditional surface-level metrics. The code used in this work is publicly available.

31 October 2025 in Preprints.org
Verified
Preprint
Open access
PDF file
RRRahul Raja
Rahul Raja
AVArpita Vats
Arpita Vats
OTOmkar Thawakar
Omkar Thawakar
TATajamul Ashraf
Tajamul Ashraf
Object tracking remains a central problem in computer vision with broad applications in surveillance, autonomous driving, augmented reality, and human–computer interaction. This paper presents a comprehensive survey that unifies the progression of tracking methodologies, from handcrafted and probabilistic models to deep learning paradigms and recent advances with large vision–language and foundation models. We categorize tracking into Single Object Tracking (SOT), Multi-Object Tracking (MOT), and Long-Term Tracking (LTT), systematically reviewing CNN, Siamese, transformer, and hybrid-based approaches alongside detection-guided, detection-integrated, and re-identification–aware pipelines. Special emphasis is placed on emerging trends, including open-vocabulary tracking, promptable models, and multimodal fusion across RGB, depth, thermal, LiDAR, and event-based inputs. We highlight benchmark datasets, evaluation protocols, and taxonomy refinements that reveal convergence toward unified and generalizable tracking systems. Finally, we discuss open challenges—such as occlusion, scalability, identity consistency, and cross-dataset transferability—and outline future directions in self-supervised learning, adapter tuning, and efficient foundation model adaptation. This survey aims to serve as a reference for both academic researchers and practitioners, bridging classical paradigms with the rapidly evolving landscape of foundation and vision- language driven tracking.