Skip to content
Log in

Author

Jeeho Ryoo

View on Semantic Scholar

2 papers indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Review Open access Aug 2026

“Should I tell my teacher?” Student AI disclosure practices, stigma, and self-regulated learning in higher education

As generative artificial intelligence (GAI) tools become popular in education, questions about when and whether students disclose their AI use have emerged as a critical concern for academic integrity and self-regulated learning (SRL). Using SRL as an interpretive lens, we examine undergraduate students’ AI disclosure practices, investigating how their worries about negative consequences, AI usage behaviours, and demographic factors relate to their willingness to disclose AI use in academic settings. Drawing on survey data from 78 undergraduate students enrolled in an online education course, we developed study-specific survey items addressing worry, disclosure, and AI use. Given the developmental status of the instrument, we retained two preliminary composite indicators, worry and disclosure to teachers, and analyzed less internally consistent peer-disclosure and AI-use items individually. Our findings reveal a pattern of co-occurring worries: fear of teacher judgment and fear of social stigmatization are strongly correlated, and both are associated with students’ active concealment of AI use. Interpretively, we describe two contrasting response patterns among students: a transparent-disclosure pattern, in which disclosing AI use to instructors appears to reflect part of SRL processes, and an anxious-concealment pattern, in which heavier AI users report less transparency with teachers. A further finding is that anxiety is not associated with reduced disclosure to teachers directly, but is associated with increased peer-only sharing and complete secrecy. Exploratory analyses also suggest demographic patterns worth further study, with academic discipline showing the largest association with disclosure and a marginal indication that monolingual students may disclose less than multilingual students. Given the sample size, we reserve our interpretations and treat these as hypotheses for future work rather than established effects. These findings suggest that institutional AI disclosure frameworks may benefit from addressing both policy compliance and the affective and disciplinary dimensions of students’ SRL decision-making.

Daniel H. Chang, M. Lin, Jing Huang et al. · 0 citations
Jul 2026

Elastic Memory Remapping for Multi-tenant LLM Serving

KV cache accelerates LLM inference by avoiding redundant computation, but its rapidly growing memory footprint makes GPU memory a primary bottleneck in modern serving systems. Recent approaches extend GPU memory using CPU memory through KV-cache swapping. However, because KV cache is continuously updated during decoding, swapping introduces substantial synchronization and bidirectional transfer overheads. We present Oneiros, a dynamic remapping engine for multi-tenant LLM serving. Oneiros is based on a simple observation: unlike KV cache, model parameters remain immutable during inference. Instead of swapping KV cache itself, Oneiros dynamically repurposes GPU memory allocated for model parameters as KV cache capacity, enabling nonblocking, unidirectional parameter transfer. This approach is particularly effective in multi-tenant environments, where memory allocated to inactive models can be reclaimed dynamically for active workloads. We implement Oneiros in vLLM and evaluate it on modern GH200 systems. Compared to vLLM, Oneiros reduces tail latency by up to 99.3% and improves throughput by up to 86.7%. Compared to KV-cache swapping approaches, Oneiros achieves substantially higher throughput by avoiding synchronization overheads during runtime memory extension. Source code of Oneiros is available at https://github.com/UT-SysML/Oneiros/1.

Ruihao Li, Shagnik Pal, Vineeth Narayan Pullu et al. · 0 citations