iiinova — where your digital self keeps evolving

Learning Accelerators @learningaccelerators.com · 19/12/2024

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:

researchsquare.com

AI Tutoring Outperforms Active Learning

Advances in generative artificial intelligence (GAI) show great potential for improving education. Yet little is known about how this new technology should be used and how effective it can be. Here we...

010

Learning Accelerators @learningaccelerators.com · 19/12/2024

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.

100

Learning Accelerators @learningaccelerators.com · 19/12/2024

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.

100

Learning Accelerators @learningaccelerators.com · 19/12/2024

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).

100

Learning Accelerators @learningaccelerators.com · 19/12/2024

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).

110

Learning Accelerators @learningaccelerators.com · 19/12/2024

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.

drphilippahardman.substack.com

AI Tutors Double Rates of Learning in Less Learning Time

Inside Harvard's new groundbreaking study

100

Learning Accelerators @learningaccelerators.com · 19/12/2024

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.

110

Learning Accelerators @learningaccelerators.com · 19/12/2024

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.

100

Learning Accelerators @learningaccelerators.com · 19/12/2024

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.

forbes.com

Students Learned Twice As Much With AI Tutor Than Typical Lectures

Harvard University researchers found AI tutored students learned more in less time but stress that AI tutors are a resource, not a replacement for teachers.

100

Learning Accelerators @learningaccelerators.com · 19/12/2024

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

news.harvard.edu

Professor tailored AI tutor to physics course. Engagement doubled. — Harvard Gazette

Preliminary findings inspire other large Harvard classes to test approach this fall

120

Learning Accelerators @learningaccelerators.com · 12/12/2024

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”

100

Learning Accelerators @learningaccelerators.com · 12/12/2024

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.”

100

Learning Accelerators @learningaccelerators.com · 12/12/2024

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.

100

Learning Accelerators @learningaccelerators.com · 12/12/2024

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.

100

Learning Accelerators @learningaccelerators.com · 12/12/2024

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.

100

Learning Accelerators @learningaccelerators.com · 12/12/2024

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.

100

Learning Accelerators @learningaccelerators.com · 12/12/2024

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.

100

Learning Accelerators @learningaccelerators.com · 12/12/2024

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.

100

Learning Accelerators @learningaccelerators.com · 12/12/2024

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

Couple walking in the countryside enjoying exercise

141

Learning Accelerators @learningaccelerators.com · 10/12/2024

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'.

110

Learning Accelerators @learningaccelerators.com · 10/12/2024

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.

110

Learning Accelerators @learningaccelerators.com · 10/12/2024

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.

110

Learning Accelerators @learningaccelerators.com · 10/12/2024

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).

110

Learning Accelerators @learningaccelerators.com · 10/12/2024

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).

110

Learning Accelerators @learningaccelerators.com · 10/12/2024

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.

110

Learning Accelerators @learningaccelerators.com · 10/12/2024

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.

110

Learning Accelerators @learningaccelerators.com · 10/12/2024

2/10 As the study authors point out, specifying requirements for virtual reality (VR) software development and documenting these is a labour-intensive process.

110

Learning Accelerators @learningaccelerators.com · 10/12/2024

9/10 In addition to the what, where and when of memory, Morales-Calva commented that individual circumstances, including cultural, personal and cognitive differences, can have a significant impact in shaping how individuals remember.

100

Learning Accelerators @learningaccelerators.com · 10/12/2024

8/10 Finally, the researchers found that when the event occurs makes a difference in what people remember. How individuals sequence events and recognise transitions between them plays a critical role. Specific events are often compartmentalised into distinct episodes to aid recall.

100

Learning Accelerators @learningaccelerators.com · 10/12/2024

7/10 What is known as spatial memory is often studied in animals, and the researchers found it is also an important aspect of what we remember. New environments command greater attention and therefore foster stronger memories when compared to familiar, routine settings.

100

Learning Accelerators @learningaccelerators.com · 10/12/2024

6/10 For example, individuals are more likely to remember events with deep emotional resonance or details on which they actively focus. However, what we remember is also influenced by factors such as where the event happened.

110

Learning Accelerators @learningaccelerators.com · 10/12/2024

5/10 The review categorizes memory research into three primary domains centered on what, where and when people remember. Morales-Calva and Leal found that memories are often shaped by emotional content, personal significance, repetition and attention.

100

Learning Accelerators @learningaccelerators.com · 10/12/2024

4/10 Specifically, the researchers explored how emotional significance, personal relevance and individual differences shape memory retention. Unlike experimental studies, this review gathers and interprets existing findings to advance the understanding of what is termed episodic memory.

100

Learning Accelerators @learningaccelerators.com · 10/12/2024

3/10 The authors Fernanda Morales-Calva and Stephanie Leal, examined existing research to create a comprehensive analysis of the "three Ws" of memory -- what, where and when we remember -- to answer the central question of why people remember.

110

Learning Accelerators @learningaccelerators.com · 10/12/2024

2/10 “Tell me why: The missing w in episodic memory's what, where and when" appears in a special issue of Cognitive, Affective & Behavioral Neuroscience focusing on individual differences in memory modulation.

100

Learning Accelerators @learningaccelerators.com · 10/12/2024

Why do people remember certain things and forget others? A recently published review paper by researchers at Rice University sheds light on the relationship between factors that influence human memory. A Thread 1/10 #cognition #memory #recall #neuroscience #research #EduSky

Woman slaps forehead in shock having realised she has forgotten something.

165

Learning Accelerators @learningaccelerators.com · 08/12/2024

10/10 Here is a link to the full evaluation study report: ‘Engaging with data using People Progress: Effects on educational leaders and learners’, published by NCFE:

ncfe.org.uk

000

Learning Accelerators @learningaccelerators.com · 08/12/2024

9/10 Other key outcomes included: • Improved student and teacher wellbeing • Cost and time savings • Enhanced communication with students, parents, and school staff. The tagging feature in Pupil Progress was found to have boost engagement significantly.

110

Learning Accelerators @learningaccelerators.com · 08/12/2024

8/10 The pilot found reliable question-level analysis data from formal assessments allowed leaders and teachers to derive accurate insights from data analysis and plan changes in course and curriculum delivery with confidence.

100

Learning Accelerators @learningaccelerators.com · 08/12/2024

7/10 The students then completed another mock exam and the change in data profiles was analysed to see correlated impact on achievement. Students, teachers and leaders gave feedback through surveys and interviews to provide insights into the granular skills data, and the impact on closing gaps.

100

Learning Accelerators @learningaccelerators.com · 08/12/2024

6/10 Teachers used the question-level analysis data from a December mock exam to inform potential changes to their curriculum, provide additional resources and target interventions. Students and parents were given reports with a breakdown highlighting areas of strength and progress opportunities.

100

Learning Accelerators @learningaccelerators.com · 08/12/2024

5/10 Each school was provided with: • Exam board-specific trackers with granular question-level data • Teacher tools to identify and tag underperforming students • Student reports with detailed skill breakdowns that helped identify progress opportunities.

100

Learning Accelerators @learningaccelerators.com · 08/12/2024

4/10 Pupil Progress is a technology platform that aims to provide an up-to-date, accurate breakdown of student performance across all subjects on a whole school level.

100

Learning Accelerators @learningaccelerators.com · 08/12/2024

3/10 This is the high stakes school year when most children aged 15 to 16 in the United Kingdom take GCSEs or other national equivalent exams.

110

Learning Accelerators @learningaccelerators.com · 08/12/2024

2/10 Thanks to a recent pilot study funded by NCFE, Year 11 English and Maths learners across 12 schools in England participated in a pilot designed to investigate the benefits of enhancing school data analysis technology.

100

Learning Accelerators @learningaccelerators.com · 08/12/2024

Pilot study involving 2,000 students shows upgrading school data analytics improves wellbeing, saves resources and boosts engagement. Thread 1/10 #schools #education #teachers #learnerdata #learneranalytics #examinations #exams #edtech #EduSky #research #gcse #KS4 #secondaryeducation

Screenshot of Pupil Progress app on mobile phone

100