Author name cluster

Ritvik Garimella

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

5 papers

1 author row

AAAI Conference 2026 System Paper

Chatsparent: An Interactive System for Detecting and Mitigating Cognitive Fatigue in LLMs

Riju Marwah
Vishal Pallagani
Ritvik Garimella
Amit Sheth

LLMs are increasingly being deployed as chatbots, but today’s interfaces offer little to no friction: users interact through seamless conversations that conceal when the model is drifting, hallucinating or failing. This lack of transparency fosters blind trust, even as models produce unstable or repetitive outputs. We introduce an interactive demo that surfaces and mitigates cognitive fatigue, a failure mode where LLMs gradually lose coherence during auto-regressive generation. Our system, Chatsparent, instruments real-time, token-level signals of fatigue, including attention-to-prompt decay, embedding drift, and entropy collapse, and visualizes them as a unified fatigue index. When fatigue thresholds are crossed, the interface allows users to activate lightweight interventions such as attention resets, entropy-regularized decoding, and self-reflection checkpoints. The demo streams live text and fatigue signals, allowing users to observe when fatigue arises, how it affects output quality, and how interventions restore stability. By turning passive chatbot interaction into an interactive diagnostic experience, our system empowers users to better understand LLM behavior while improving reliability at inference time.

PDF Details DOI

AAAI Conference 2026 Conference Paper

DETONATE – A Benchmark for Text-to-Image Alignment and Kernelized Direct Preference Optimization

Renjith Prasad Kaippilly Mana
Abhilekh Borah
Hasnat Md Abdullah
Chathurangi Shyalika
Gurpreet Singh
Ritvik Garimella
Rajarshi Roy
Harshul Raj Surana

Alignment is crucial for text-to-image (T2I) models to ensure that the generated images faithfully capture user intent while maintaining safety and fairness. Direct Preference Optimization (DPO) has emerged as a key alignment technique for large language models (LLMs), and its influence is now extending to T2I systems. This paper introduces DPO-Kernels for T2I models, a novel extension of DPO that enhances alignment across three key dimensions: (i) Hybrid Loss, which integrates embedding-based objectives with the traditional probability-based loss to improve optimization; (ii) Kernelized Representations, leveraging Radial Basis Function (RBF), Polynomial, and Wavelet kernels to enable richer feature transformations, ensuring better separation between safe and unsafe inputs; and (iii) Divergence Selection, expanding beyond DPO’s default Kullback–Leibler (KL) regularizer by incorporating alternative divergence measures such as Wasserstein and Rényi divergences to enhance stability and robustness in alignment training. We introduce DETONATE, the first large-scale benchmark of its kind, comprising approximately 100K curated image pairs, categorized as chosen and rejected. This benchmark encapsulates three critical axes of social bias and discrimination: Race, Gender, and Disability. The prompts are sourced from the hate speech datasets, while the images are generated using state-of-the-art T2I models, including Stable Diffusion 3.5 Large (SD-3.5), Stable Diffusion XL (SD-XL), and Midjourney. Furthermore, to evaluate alignment beyond surface metrics, we introduce the Alignment Quality Index (AQI) for T2I systems: a novel geometric measure that quantifies latent space separability of safe/unsafe image activations, revealing hidden model vulnerabilities. While alignment techniques often risk overfitting, we empirically demonstrate that DPO-Kernels preserve strong generalization bounds using the theory of Heavy-Tailed Self-Regularization (HT-SR).

PDF Details DOI

AAAI Conference 2026 System Paper

In-Situ Eval: A Modular Framework for Custom and Real-Time RAG Benchmarking

Ritvik Garimella
Kaushik Roy
Chathurangi Shyalika
Amit Sheth

Retrieval-Augmented Generation (RAG) has become the standard approach for integrating domain knowledge into Large Language Models (LLMs). However, fair comparison of RAG pipelines remains difficult: data preparation is often ad hoc, subsampling methods are opaque, parameters vary across implementations, and evaluation is fragmented. We present In-Situ Eval, a unified and reproducible framework that operationalizes the full RAG pipeline with configurable subsampling strategies and both RAG-specific and generic evaluation metrics. The platform supports two execution modes: an offline Dataset mode for evaluating precomputed outputs, and a live Retrieval mode for benchmarking RAG variants with state-of-the-art LLMs. Users can flexibly select datasets, retrieval techniques, models, and metrics, enabling side-by-side comparisons, ablations, and targeted analyses. This holistic approach reduces computational costs, clarifies the impact of subsampling techniques, and provides actionable insights for real-world deployments. By facilitating transparent, customizable, and interactive benchmarking, In-Situ Eval empowers both researchers and practitioners to make informed decisions in adapting RAG pipelines to domain-specific needs.

PDF Details DOI

AAMAS Conference 2026 Conference Paper

KiMO: Knowledge-infused Multi-agent Orchestrator

Ritvik Garimella
Michael Charles Stewart
Kevin Do
Leeon Israel
Amit Sheth

Multi-Agent Systems (MAS) provide a natural paradigm for solving complex problems requiring coordinated, multi-step reasoning and decision making. As these systems scale, effective agentic orchestration becomes crucial. While existing tools such as Agent- Flow and Agent Development Kit (ADK) support agent communication and task delegation, they remain limited in understanding task semantics, coordinating heterogeneous agent types (e. g. , reactive vs. cognitive), and aligning agent selection with domain knowledge and infrastructure constraints. We present KiMO, a knowledge-infused multi-agent orchestrator that makes coordination explicit through structured knowledge infusion: planning ontologies encode domain-specific task decompositions, and agent registries formalize capabilities and compatibility constraints for heterogeneous components. KiMO operationalizes these through a two-stage pipeline: ontology-guided sub-task planning (Plan- Gen)andregistry-guidedworkflowassembly(AgentGen). Thisdual knowledge infusion enables interpretable, flexible orchestration with human-inspectable coordination decisions. Our interactive platform demonstrates real-time workflow construction for manufacturing pipelines with expert refinement and dynamic knowledge extension. Code is available at https: //tinyurl. com/3es9zyhj and Demo Video at https: //tinyurl. com/yfa8ctwn.

PDF

AAMAS Conference 2026 Conference Paper

Neurosymbolic Framework for building Robust, Efficient, and Explainable Multi-Agent Enterprise Systems

Ritvik Garimella

Multi-agentsystemsforenterpriseautomationisanintelligentsolutionforaubiquitousproblemwiththreemainobjectives, robustness, efficiency and explainability. First, we study robustness in terms of automation procedures for similar tasks by aiming towards a construction of a hierarchical task decomposition with dynamic agent management system. Secondly, we study the concept of efficiency through reinforcement learning based algorithms for fine-tuned refinementsofagentutilizationandparametricoptimization. Lastly, we investigate the concept of explainability by building evaluation suites for agents evaluation on both syntactic and semantic levels. The expected contributions include theoretical foundations for efficient orchestration and practical evaluation tools, advancing deployment of production-ready multi-agent systems.

PDF