Shivani Kumar
Papers
1
Total Citations
2
H-Index
1
About
Shivani Kumar investigates the intersection of natural language processing and human-computer interaction, with a focus on dialogue systems and the evaluation of large language models (LLMs). Her work addresses critical challenges in simulating human-like conversational behavior, particularly the trade-offs between cost-effective automation and authentic human response qualities. In her highly cited 2024 paper, "Real or Robotic? Assessing Whether LLMs Accurately Simulate Qualities of Human Responses in Dialogue," Kumar systematically evaluates how well LLMs replicate nuanced human dialogue traits—such as empathy, coherence, and spontaneity—in simulated interactions. This research has significant implications for reducing the expense and time required to build dialogue datasets, traditionally reliant on recruiting and training human participants. By demonstrating both the potential and limitations of LLM-based simulation, Kumar provides a framework for more efficient data collection while cautioning against over-reliance on synthetic responses. Her contributions are gaining traction among researchers developing conversational AI, with her work cited in studies on dialogue system evaluation and human-LLM interaction modeling. Kumar’s findings help shape best practices for using LLMs as proxies in behavioral research, advancing the field toward more scalable yet trustworthy dialogue technologies.
Research Focus
Key Achievements
Top Papers
- 1