Data Science and AI Educators' Programme: Cohort Call Two notes

Data Science and AI Educators' Programme: Cohort Call Two notes === ###### tags: `cohort-call-2` `DS-AI-Educators'-Programme` :::info - **Call time and day**: Wednesday 17 May 2023, 13:30 - 15:00 (GMT+1) - **Meeting host**: Ayesha - **Panel chair**: Matt - **Call joining link**: - **Github repo**: [DS and AI Educators' Programme GitHub](https://github.com/alan-turing-institute/ds-ai-educators-programme) ::: ## CC2: Challenges of teaching Data Science and AI ### Agenda | Item | Who | Timing | | ---------------------------- | ------------- | ------ | | Intro and housekeeping | Ayesha Magill | 13:30 - 13:40 | | Breakout rooms - what challenges do you face in your teaching of DS and AI? | All | 13:40 - 13:55 | | Panel, inc Q&A | Matt Forshaw, The Alan Turing Institute </br> [David Stern, IDEMS International](https://www.idems.international/our-team/) </br> [Graham Cole, Newcastle University](https://www.linkedin.com/in/graham-cole-8767a41b6/?originalSubdomain=uk) </br> [Ogerta Elezaj, Birmingham City University](https://www.linkedin.com/in/ogerta-elezaj-71251115/) | 13:55 - 14:45 | | Resource sharing | All | 14:45 - 14:55 | | Wrap-up and closing remarks | Ayesha Magill | 14:55 - 15:00 | ### Questions for panelists Introductions: - A little bit about you, who you are and where you're joining from. - What is your role? - How does Data Science and AI feature in your role? ### Questions from participants Place your questions here and we will pick these up in the panel session :smile_cat: - Where do you see AI/DS education going? - How to balance domain-specific knowledge with DS and AI skills? Or, how much domain-specific knowledge do students need to be effective Data Scientists in that domain? - Do you have a predefined set of learning objectives/ concepts/ competencies to be learned and an underpinning pedagogical approach? With a view of what the mental models are/ progression of the notational machine for ML systems? Are there curated data sets that are available that can be used to teach particular concepts/ competencies? - How do we make the most of the course readings? How much should we as the teacher count on students reading the course materials. Good books often do an amazing job at teaching the subject and more of the course can then be used on discussion and answering questions. - How do you help your students orientate around DS & AI careers given the diversity range of options, roles, and how fast moving these are? - How can we create a climate whereby resources are shared for the teaching and learning of DS/AI? What barriers do you think there are to sharing resources? IP/ competition/ nervousness that resources are not "good enough" ?? - - - ### Useful Links/references: - AI and data literacy education research from the [Computing Education Research Centre](https://computingeducationresearch.org/) - [Artificial intelligence, machine learning and data science seminar series (Sep 2021 – Mar 2022)](https://www.raspberrypi.org/computing-education-research-online-seminars/previous-seminars/) - [Turing Research Software Engineering Course](https://alan-turing-institute.github.io/rse-course/) - [Turing Research Data Science Course](https://alan-turing-institute.github.io/rds-course/) - [How data lies](https://www.turing.ac.uk/courses/how-data-lies) course - [Data Science Alliance competency framework](https://rss.org.uk/membership/professional-development/data-science-standards/) - [Data8 - conference and open teaching resource in statistical and inferential thinking](http://www.data8.org) - [AI fairness on social media](https://learn.turing.ac.uk/course/view.php?id=20) - you will need to register for a free account to access this - [Journal of Open Source Education](https://jose.theoj.org/about) - you might find useful articles or be able to publish some of your own resources! - [Data Carpentries lessons](https://datacarpentry.org/lessons/) - [Kaggle Learn](https://www.kaggle.com/learn) - useful set of resources/notebooks on topics such as machine learning, SQL etc. - [The Box](https://aiethicslab.com/the-box) - AI ethics toolbox - [ISCB Competency framework](https://competency.ebi.ac.uk/framework/iscb/3.0/competencies) - [Turing Commons](https://alan-turing-institute.github.io/turing-commons/) - resources for training in AI ethics and governance and responsible research and innovation. - [MIT Primary School AI Education](https://www.media.mit.edu/projects/primary-ai-ed/overview/) ### Common themes from breakout rooms - Little experience in teaching - [name=Ayesha] ... - Finding time for each student - [name=Matt] I have always found it a very difficult balance. I want to be maximally available to learners, and to be available in whichever way is most helpful to learners (e.g. timetabled lab sessions, email, Teams chat) but also to try and encourage learners to come together in a critical mass at sessions to support peer-to-peer support. When I reflect on my previous practice, I probably caused situations where I was quite overworked by handling a lot of queries 1:1, but I always found it quite difficult to take a more stern line. I would probably find myself doing things the same way again :smile_cat: - [name=Wairimu] - especially when there are many learners and other competing demands on time - [name=Ann] - not always possible, but in large classes I've found it key to get additional support with this (e.g. peer-peer learning, in a university setting getting students to help as teaching assistants, FAQ drop-in session etc). As Matt says, trying to provide all the help yourself on a 1:1 basis definitely risks being too overwhelming! - Challenge of embracing the 'new' - [name=Matt] I would attempt to create an environment of shared learning and 'mutual struggle' but learners' may often expect the trainer to have all the answers. I find live coding and approaches similar to this are a good way to humanise the process. - Fast-changing nature of AI/DS (inc. evolving curriculum so needs updating) - [name=Wairimu] - especially due to having to update for each domain - Teaching students from non STEM (maths anxiety) - [name=Cari] working with students from a law background, many do not have any experience in STEM. Some may not have entry-level awareness of terminology and concepts, even if they show initial enthusiasm care needs to be taken not to put them off. - [name=Andrea] - balancing different student backgrounds and expectations is difficult when you have to teach blanket courses with generalist outcomes. I also sturggle with some impostor syndrome, which is not the best quality of an educator. However, todays discussion, especially David's comments on application based teaching gave me a boost. - rest in learning DS , AI or programming - Lack of learner's prior knowledge and therefore differentiating - [name= Joyce] we prepared lesson plans and color coded activities to aid with differentiation - Access to the right tools for teaching [name=Wairimu] when teaching online and when learners are not willing to learn new tools (Zoom fatigue is real!) - [name= Joyce] we only used open sources software - - Difficulty with online learners - how can we assess, support etc? - Convincing both students that the skills are relevant and also professionals that we need upskilling - [name=Cari] professional lawyers / judges - some will not see the importance of a more nuanced awareness in this subject (may react in a binary way - all good or all bad) - Managing media narratives around DS and AI ### Breakout room discussions #### Room 1 _What challenges do you face, in your role as an educator, when teaching data science and/or AI?_ - Not much experience in teaching; students would have different background knowledge (esp in coding and others have not much) --> will be difficult for lecturer to find starting point to approach the subject - Challenge to give each student value of their time - How to make the lesson have more value than watching videos on youtube or online courses - Challenge for new: fear of the unknown; - Weather - Identity as a teacher affects student reception - How to engage students with different learning preferences _What challenges, if any, have others within your field encountered?_ - - - - - - - - - _What solutions, if any, have you implemented to help overcome these challenges? Please give details._ - Set expectations based on the learning objectives - - - - - - - - --- #### Room 2 _What challenges do you face, in your role as an educator, when teaching data science and/or AI?_ - Having a cohort of mixed background - - fast changing nature of the AI/DS - - - - - - _What challenges, if any, have others within your field encountered?_ - The expectation of learners that they will have sufficient knowledge by the end of the programme to get into corporate space of practical application - - - Keeping students motivated through hard tasks/topics” - - - - _What solutions, if any, have you implemented to help overcome these challenges? Please give details._ - Adjust the learner’s expectations - - Having innovate practical tasks to motivate learners - - more case studies related to the students’ background - - - - --- #### Room 3 _What challenges do you face, in your role as an educator, when teaching data science and/or AI?_ - teaching students from non-STEM Backgrounds as they lack the confidence not the skills - tailor content to different audeicnes - build confidence and understanding outside technical skills eg critical thinking - - - - - - _What challenges, if any, have others within your field encountered?_ - - - - - - - - - _What solutions, if any, have you implemented to help overcome these challenges? Please give details._ - diversity teaching staff - bring together students from different backgrounds eg ethics and CS - contextualise training with project and group work - - - - - - --- #### Room 4 _What challenges do you face, in your role as an educator, when teaching data science and/or AI?_ - Access to right tools for teaching - No ownership over what people are learning- subscription to outsourced platform - Only using open source software in a recent course. - Lack of prior knowledge from learner - Differentiation can be challenging - - - - _What challenges, if any, have others within your field encountered?_ - People who can affect the flow of the session by coming off of mute and mentioning something that is either beyond the scope of the course or will come up later. - Wide variety of applications means not much is universally relevant - - - - - - - _What solutions, if any, have you implemented to help overcome these challenges? Please give details._ - 1:1 coaching for motivated learners - Techlist - Time at the beginning of a course to ensure people are set up correctly. - Provide a drop in session prior to the course for any edge case technical issues. - Recording sessions for learners to revisit. - - - - --- #### Room 5 _What challenges do you face, in your role as an educator, when teaching data science and/or AI?_ - Learners from different background - Motivate learners - Convince them the skills are useful - Learners with no background in STEM - maths anxiety - Difficulty with online courses - uncertainty - how are learners progressing? And how to support students when teaching asynchronously online so can't answer questions immediately. - How to convince busy professionals to make time for these skills - DOn't have relevant experience myself always - - _What challenges, if any, have others within your field encountered?_ - lack of engagement with online course (self-motivation etc) - - - - - - - - _What solutions, if any, have you implemented to help overcome these challenges? Please give details._ - using metaphors and innovative methods to explain building block concepts - getting faster learners to help those who are struggling more, both benefit - let students progress at own pace - show them how a technique is used before teaching the concept (to help show it's useful) - If professional /otherwise reluctant - ensure they know how this skill will benefit them! - Ensure students know where to go for further information - - - --- #### Room 6 _What challenges do you face, in your role as an educator, when teaching data science and/or AI?_ - Agricultural sciences: Perceptions around issues of working with data (undergraduates studying agriculture, farmers)... Most dont' feel like they have a robust interest in data science, or feel like they're not good at (which can make it a self-fulfilling prophecy) - Biomedical sciences: People who don't feel like this is not their domain expertise, building confidence in data science, maths - Energy storage systems: Teaching skills for projects, people are from wide range of backgrounds, they don't have prior knowledge about AI or data-driven methods, managing expectations, keeping them motivated. Not just output oriented - Anthropology->DS: Fields that use DS and AI should also contribute to DS and AI (the flow of knowledge seems mainly to be the other way, which may contribute to motivation and involvement in this shift) - Don't have foundation in basic IT skills, in limited time, is used for basic IT skills - Had to convince colleagues that this should be a priority skillset - Educators tend to get categorised as the one-stop-shop for all things IT, data science, etc. - Classes can be modular, rather than a regular integration into school curriculum - Circular process: where people are not motivated or feel like it is difficult to catch up, become - Often IT & Tech education can feel very top down ("we have to do this"), which can often decrease the desire to learn - _What challenges, if any, have others within your field encountered?_ - - - - - - - - - _What solutions, if any, have you implemented to help overcome these challenges? Please give details._ - Playful applications of knowledge from trainings: making Github pages website, making Github PR, Digital Gardens - https://digital-garden-jekyll-template.netlify.app/ - - - - - - - - --- #### Room 7 _What challenges do you face, in your role as an educator, when teaching data science and/or AI?_ - Motivation re: quant skills when not from a quant background - Bad experiences with Maths previously - Press narratives and misunderstandings of technology - Unfamiliarity with coding or thinking in a data science/AI way - Wanting to implement AI/DS without understanding why and implications of doing so - - - - _What challenges, if any, have others within your field encountered?_ - Changes in technology - Platforms to teach with and meaningful data - - - - - - - _What solutions, if any, have you implemented to help overcome these challenges? Please give details._ - Relevant examples and data sets - Data carpentries lessons - - - - - - - --- #### Room 8 _What challenges do you face, in your role as an educator, when teaching data science and/or AI?_ - data science combines many fields (math, statistics, programming) - terminology, syntax, grammer first without knowing full implications - evolving field, so need to update curriculum often - balance between theory and hands-on - data science has an element of art - - - - _What challenges, if any, have others within your field encountered?_ - - - - - - - - - _What solutions, if any, have you implemented to help overcome these challenges? Please give details._ - include hands on exercises - update curriculum - include ethical components in teaching - - - - - - --- #### Room 9 _What challenges do you face, in your role as an educator, when teaching data science and/or AI?_ - wide range of student backgrounds - lack of time - assessments? especially facing chatGPT and LLMs? maybe we can ask students to explain. If copilot suggests something, ask them to critic. Whatis they use LLM to critic? - what we should teach? in what order? and how? - everyone is nervous/traumatized to do show and tell especially when you are new - some resources are IP restrictions and we cannot share - - - _What challenges, if any, have others within your field encountered?_ - - - - - - - - - _What solutions, if any, have you implemented to help overcome these challenges? Please give details._ - clear learning objectives - assign TAs to each 10 students - share/open-source materials and learning objectives if possible - - - - - - --- #### Room 10 _What challenges do you face, in your role as an educator, when teaching data science and/or AI?_ - When teaching a problem solving course in data science to a pan African doctoral training school in data science I was confronted with the lack of data exposure as most participant were very methods oriented with their skills and experience. - My biggest challange is accommodating the different levels and backgrounds of my students, as I teach data science for both STEM and Humanities students. - An other challange of mine is my own psychological barriers, as I am not coming from a Computer Science background, but am self-trained in data science, and has a humanities-social sciences educational background. - I find it challanging keeping up with with the new projects and the developments in the field while designing and renewing course materials. - Our program is interdisciplinary, and find co-faculty to teach courses with ethics/ social justice at the core has been tough. - Teaching professionals from different backgrounds; so making sure we customise learnings for different fields. This is not always easy to do - Teaching students from a CIS background to unlearn bad habits has been tough - Teaching online has also been a challenge especially keeping learners engaged and sustaining their motivation. - - - - - _What challenges, if any, have others within your field encountered?_ - - - - - - - - - _What solutions, if any, have you implemented to help overcome these challenges? Please give details._ - In response to the lack of data experience I actually have found that complex simulated data enables me to expose them to gaps in their understanding. - Project and case study based classes can engage students with diverse background in solving one shared problem - - - - - -

Syntax	Example	Reference
# Header	Header	基本排版
- Unordered List	Unordered List
1. Ordered List	Ordered List
- [ ] Todo List	Todo List
> Blockquote	Blockquote
Bold font	Bold font
Italics font	Italics font
~~Strikethrough~~	~~Strikethrough~~
19^th^	19^th
H~2~O	H₂O
++Inserted text++	Inserted text
==Marked text==	Marked text
[link text](https:// "title")	Link
![image alt](https:// "title")	Image
`Code`	`Code`	在筆記中貼入程式碼
```javascript var i = 0; ```	`var i = 0;`
:smile:		Emoji list
{%youtube youtube_id %}	Externals
$L^aT_eX$	L^aT_eX
:::info This is a alert area. :::	This is a alert area.