Article Per Year (5 Year)
p-Index From 2021 - 2026
0.61
P-Index
This Author published in this journals
All Journal Journal of Technology Informatics and Engineering
Xiaohan Chang
Computer Science, University of Connecticut, CT, USA
Computer Science & IT

Published : 3 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 3 Documents
Search

 1 
Distilling VMAF into an Edge-Deployable Quality Predictor: A Pilot Shot-Level Proxy with LLM-Ready Quality Tokens Binghua Zhou; Heyu Wang; Xiaohan Chang
Journal of Technology Informatics and Engineering Vol. 4 No. 2 (2025): AUGUST | JTIE : Journal of Technology Informatics and Engineering
Publisher : University of Science and Computer Technology

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.51903/jtie.v4i2.522

Abstract

This pilot study evaluates whether a compact student model can approximate VMAF well enough to support low-latency release guarding on edge-class CPU environments. The corpus comprises a 62.31-second Big Buck Bunny excerpt at 1280 × 720 and 25 fps, segmented into 13 shots. Twelve distorted variants were generated by crossing H.264/AVC and H.265/HEVC with 180p, 240p, and 360p delivery resolutions and two quality levels per codec-resolution pair, yielding 156 shot-level samples. Frame-level VMAF scores were aggregated into shot-level teacher labels, and a student proxy consumed 14 low-cost no-reference features derived from decoded frames and stream metadata. Shot-grouped five-fold cross-validation was used to prevent content leakage across train-test splits. On this corpus, a 50-tree gradient-boosted decision tree achieved MAE = 6.56 VMAF points, RMSE = 8.32, and Pearson r = 0.913. Relative to simple regressors, the student reduced MAE by approximately 21.5% versus bitrate-only regression and 10.7% versus metadata-only regression. In a single CPU-only benchmark, predictor latency was 0.484 ms per sample and the full decode-feature-predict chain averaged 42.61 ms versus 1117.41 ms for the teacher, corresponding to a 26.22× end-to-end speed-up. As a thresholded guard, the same student reached F1 = 0.826, 0.893, and 0.900 at 60, 70, and 80 VMAF respectively. These findings support the feasibility of a practical edge proxy on this specific pilot corpus, but they should not be interpreted as broad generalization across content classes or production ladders. The paper also introduces an LLM-ready token interface intended for downstream reporting rather than for replacing the underlying quality measurement
Layout-Aware Progressive PDF Rendering: AI Prioritization of PDF Slices to Reduce Time-to-Functional-First-Frame on FUNSD Heyu Wang; Yuxuan Ren; Xiaohan Chang
Journal of Technology Informatics and Engineering Vol. 4 No. 2 (2025): AUGUST | JTIE : Journal of Technology Informatics and Engineering
Publisher : University of Science and Computer Technology

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.51903/jtie.v4i2.523

Abstract

Progressive PDF rendering is attractive because users rarely need every visible pixel at once; they need the semantically useful parts of the current viewport early enough for reading and interaction. This paper studies whether layout-aware AI can prioritize PDF slices more effectively than geometric or density-based heuristics. We reconstruct vector PDFs from official FUNSD form annotations and evaluate a tile scheduler that predicts tile utility from inexpensive layout and preview features before high-resolution rendering begins. The empirical study covers 26 reconstructed documents from the FUNSD test split that were fully processed in the present environment, four viewport scenarios, and measured clip-render timings for all visible tiles. The main configuration uses an 8×10 grid and a random-forest regressor trained with page-level 5-fold GroupKFold, then compares the learned scheduler with row-major visible-first, center-first, ink-density, text-density, a hand-tuned layout heuristic, full-page rendering, and an oracle upper bound. The proposed model reaches TTFF-90 in 14.21 ms, compared with 15.18 ms for the best non-AI heuristic, 20.48 ms for full-page rendering, and 24.09 ms for row-major rendering. It also achieves Utility@20ms of 0.941, AUC@25ms of 0.730, NDCG@10 of 0.963, and Recall@10 of 0.969. The results show that slice rendering is not inherently beneficial: the summed visible-tile cost in the main 8×10 setting is 28.80 ms, which is higher than the full-page cost of 20.48 ms, so scheduling quality determines whether slicing improves or harms TTFF. A coarser 6×8 grid reduces AI TTFF-90 to 10.58 ms, while the densest pages favor a full-page fallback. Paired Wilcoxon signed-rank tests over the page-scenario cases yield p < .001 for TTFF-90 improvements of the proposed model over every non-AI baseline. However, those tests should be interpreted as case-level rather than document-level evidence.
Profit-Aware Spot GPU Admission Control with Cost-Sensitive Loss and Evidence-Grounded Policy Memos for AI Workload Supply-Demand Matching Siming Zhao; Yuxuan Ren; Xiaohan Chang
Journal of Technology Informatics and Engineering Vol. 5 No. 2 (2026): AUGUST | JTIE : Journal of Technology Informatics and Engineering
Publisher : University of Science and Computer Technology

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.51903/jtie.v5i2.545

Abstract

AI clusters increasingly operate with heterogeneous GPU resources where production workloads and opportunistic spot jobs compete for limited accelerator capacity. This study presents a trace-driven admission-control framework using the Alibaba cluster-trace-v2026-spot-gpu dataset, consisting of 466,867 job records and 4,278 GPU-node records. The experiment evaluates GPU demand forecasting, profit-aware spot admission control, and evidence-grounded operational policy generation using chronological training, validation, and test splits. Hourly spot GPU demand forecasting was evaluated across six GPU models, where Ridge regression achieved the best test performance with an RMSE of 38.50 requested GPUs per hour, improving over both last-hour and seasonal naive baselines. The admission-control evaluation compared FIFO, greedy packing, classifier-based acceptance, utility ranking, and the proposed cost-sensitive policy. The proposed approach achieved a test profit of 67,278.96, improving 1.97% over the accuracy-oriented classifier while increasing spot success rate and reducing costly false acceptances by 13.17%. Sensitivity analysis showed that the optimal policy depends on the protection cost assigned to high-priority workloads. A deterministic evidence-grounded explanation layer generated 500 policy memos and passed numeric, policy, and evidence consistency checks. The findings suggest that profit-aware admission control can serve as a practical scheduling guardrail before detailed GPU placement and resource allocation decisions.
Co-Authors Binghua Zhou Heyu Wang Heyu Wang Siming Zhao Yuxuan Ren Yuxuan Ren