๊ฐ€์ž… ํ›„ ์ดˆ๋Œ€ ๋งํฌ๋ฅผ ๊ณต์œ ํ•˜๋ฉด ๋™์˜์ƒ ์žฌ์ƒ ๋ฐ ์ดˆ๋Œ€ ๋ณด์ƒ์„ ๋ฐ›์„ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

cv usk
@cv_usk
AI / Software Research Notes AI Agent, LLMOps, MLOps, Software Architecture ๆŠ•็จฟใฏๅ€‹ไบบใฎๆ„่ฆ‹ใงใ™ใ€‚
๊ฐ€์ž… May 2026
282 ํŒ”๋กœ์ž‰ ์ค‘    431 ํŒฌ
๐Ÿ“š Can an AI tutor help a student learn as much as a real human tutor does? A study with 2,469 participants just answered that. Title: StudentBench: AI and human tutoring yield equivalent GRE learning gains URL: โ“ What did they compare? For GRE quantitative and verbal reasoning, they compared three conditions - AI tutoring, tutoring by professional human tutors, and watching an educational video - measuring the test-score gain before and after a one-hour session. โ“ What did they find? On the combined score, the gap between AI and human tutoring was just -0.58 percentage points, statistically equivalent (p=.015). Six individual models, including Claude Opus and Gemini 3.5 Flash, each matched human tutors on their own. โ“ How much cheaper is it? The most cost-efficient model, Gemma 4 31B, matched human-level learning gains at 1/918th the cost per percentage point of learning gain. โ“ Where do humans still win? Human tutors still averaged the best results across the three verbal reasoning areas, and the study also found a causal chain where faster AI responses led to more student messages, more correct practice, and ultimately bigger learning gains. #AIEducation# #LLM#
๋” ๋ณด๊ธฐ