10/10
Here is the preprint version of the study report with citation:
Gregory Kestin*, Kelly Miller*, Anna Klales et al. AI Tutoring Outperforms Active Learning, 14 May 2024, PREPRINT (Version 1) available at Research Square:
@learningaccelerators.com.bsky.social
We help training, education, and learning technology providers accelerate learning – shaping the workforce of tomorrow, today! www.learningaccelerators.com
10/10
Here is the preprint version of the study report with citation:
Gregory Kestin*, Kelly Miller*, Anna Klales et al. AI Tutoring Outperforms Active Learning, 14 May 2024, PREPRINT (Version 1) available at Research Square:
9/10
Inspired by these results, the Derek Bok Center for Teaching and Learning is collaborating with Harvard University Information Technology to pilot similar AI chatbots in other courses. They are also developing resources to enable any instructor to integrate tutor bots into their programmes.
8/10
3. Greater Efficiency: AI sessions were shorter (median 49 minutes vs. 60 minutes) while achieving better outcomes, with 70% finishing within 60 minutes.
7/10
2. Higher Engagement and Motivation: AI users reported greater engagement (4.1/5 vs. 3.6/5) and motivation (3.4/5 vs. 3.1/5).
6/10
Key study findings include:
1. Enhanced Learning Gains:
Students using the AI tutor achieved more than double the learning gains of the classroom group, with highly significant statistical results (z = -5.6, p < 10⁻⁸).
5/10
Controls aimed to ensure fairness, including matched demographics, unbiased test designs, and separate instructors.
Dr Philippa Hardman has completed this detailed review of the study exploring the study's methodology, for those seeking independent insights.
4/10
The AI group used “PS2 Pal,” powered by GPT-4, which provided structured, self-paced guidance with pre-recorded videos and interactive support.
The classroom group attended 75-minute in-person sessions with peer discussions and instructor guidance.
3/10
The study trailed in the Harvard Gazette in September compared AI tutoring to active classroom learning using a crossover design. Students studied two topics, surface tension and fluid flow, in both settings.
2/10
The study was led by lecturer Gregory Kestin and senior lecturer Kelly Miller who analysed the learning outcomes of 194 students enrolled in autumn 2023 in Kestin’s Physical Sciences 2 course, which is physics for life sciences majors.
This Forbes article explains the circumstances.
A recent Harvard University study, currently under peer review, finds providing students with an AI tutor tailored on pedagogical principles doubled engagement and learning gains in less time than traditional face-to-face sessions.
Thread 1/10
#physics #pedagogy #ai #genai #EduSky #universities
10/10
Follow the link to the full research report: “Associations of accelerometer-measured sedentary behavior and physical activity with sleep in older adults" below.
9/10
Getting more sleep, particularly deep sleep, seems to add to this memory improvement.
"Moderate or vigorous activity means anything that gets your heart rate up -- this could be brisk walking, dancing or walking up a few flights of stairs. It doesn't have to be structured exercise”
8/10
Lead author Dr Mikaela Bloomberg (said: "Our findings suggest that the short-term memory benefits of physical activity may last longer than previously thought, possibly to the next day instead of just the few hours after exercise.”
7/10
They also accounted for participants' average levels of activity and sleep quality across the eight days they were tracked, as participants who are habitually more active and typically have higher-quality sleep perform better in cognitive tests.
6/10
They also quantified sleep duration and time spent in lighter (rapid eye movement, or REM) sleep and deeper, slow-wave sleep. The team adjusted for a wide variety of factors that might have distorted the results.
5/10
The research team looked at data from 76 men and women who wore activity trackers for eight days and took cognitive tests each day.The data was used to determine how much time participants spent being sedentary, doing light physical activity, and doing moderate or vigorous physical activity.
4/10
Less time spent sitting and six hours or more of sleep were also linked to better scores in memory tests the next day.
More deep (slow-wave*) sleep also contributed to memory function. This accounted for a small portion of the link between exercise and better next-day memory.
3/10
A new study, just published in the International Journal of Behavioral Nutrition and Physical Activity, found that, on average, people aged 50 to 83 who did more moderate to vigorous physical activity than usual on a given day did better in memory tests the day after.
2/10
Previous research in a laboratory setting has shown that people's cognitive performance improves in the hours after exercise, but how long this benefit lasts was unknown.
Couple walking in the countryside enjoying exercise
The cognitive boost our brains get from exercise persists throughout the following day, a new study by researchers from University College London suggests!
A thread - 1/10
#cognition #neuroscience #memory #psychology #physiology #exercise
10/10
Karre Sai Anirudh , Reddy Y. Raghu, Model-based approach for specifying requirements of virtual reality software products.
Frontiers in Virtual Reality 5, 2024
9/10
There is a lot more interesting detail in the full research paper by Dr. Sai Anirudh Karre, PhD and Y. Raghu Reddy, below - freshly published in the journal 'Frontiers in VR'.
8/10
VR authoring tools focus on clarity and conciseness, whilst requirement specification tools are designed to provide a more comprehensive approach to the management of requirements.
7/10
Whilst VR authoring tools are designed for developing and organising content based on requirements, requirement specification tools are an evolving class of products for managing and tracking requirements throughout development.
6/10
This uses a model-based approach to the specification of virtual reality software products which allows analysts to specify requirements for both simple and complex multi-scene VR software products and virtual environments (VEs).
5/10
Since VR related technologies are constantly evolving, VR practitioners need assistance specifying non-volatile requirements for a minimum viable VR software product.
To meet this need the study authors have developed a virtual reality requirement specification tool (VReqST).
4/10
Most VR practitioners rely on conventional software engineering (SE) requirement-gathering techniques, but the researchers identified a need for innovative methods in order to streamline VR software development.
3/10
This involves compiling detailed information on scene flow, itemising the items featured in each scene, action responses, custom behaviors, and timeline of events. The slightest change in requirements could escalate design and development costs.
2/10
As the study authors point out, specifying requirements for virtual reality (VR) software development and documenting these is a labour-intensive process.