18 November 2020 Minutes
Minutes for the Online Informatics Board of Studies Meeting at 14:00hrs-15:00hrs, Wednesday 18th November 2020, Blackboard Collaborate
Present: Björn Franke (Convenor), Sharon Goldwater (Chair), Barbara Kulawik (Administrative Secretary), Stuart Anderson, Gillian Bell, Bob Fisher, James Garforth, Paolo Guagliardo, Michael Hermann, Ava Khamseh, Nadin Kokciyan, Oisin Mac Aodha, Iain Murray, Perdita Stevens, Raul Garcia-Patron Sanchez [list inclusive but not exhaustive due to online meeting arrangement]
Apologies: Angela Nicholson, Arno Onken, Subramanian (Ram) Ramamoorthy, Steve Tonnau, Philip Wadler, Chris Williams
Minutes of Previous Meeting - 15th July 2020 (Online Consultation Only)
The Board approved all of the updates for AY 20-21 only, following online consultation. Course organizers who wish to keep the changes for future years will need to reapply to BoS later in the academic year (date TBD). https://uoe.sharepoint.com/:x:/s/InformaticsTeachingSupport/EWx7Hb6d_MJHlroL5EOpCQkBK1dM97zQTN9Ag-VmTDi6qQ?e=RYZfjC
Welcome to all members
AOCB: Change of procedure in relation to BoS operating model was taken first by Björn Franke (BF).
BF informed the Board about a change of procedure in relation to BoS operating method to bring it in line with the University regulations. The following updates apply:
- With immediate effect, the BoS is to focus on resource considerations for courses and programmes, including updates and new course proposals.
- Statements on matters such as teaching support and physical space resources will be requested with a view to discussing them more broadly in BoS meetings
- Head of School or a delegate will be asked to feed into this process
18112020_BoS Item 1 - Minor change to the MSc in Data Science - Paolo Guagliardo
Paolo Guagliardo (PG) noted that some misleading information was found in DPTs for MSc in Data Science, namely the course MATH11176 - Statistical Programming was listed as one of a Group A1 options open to taught PG students in Machine Learning, Statistics, and Optimization, yet it appeared to be available to Mathematics MSc students only. The proposed resolution was to remove the course from the DPT.
BF has asked if anyone has been in contact with Mathematics to establish whether or not this has been an intended restriction. PG confirmed that there has been an exchange of emails with Colin Rundel in Mathematics yet no confirmation has been received. PG proposed contacting the Course Organiser to check whether this could have been in fact an unintended restriction. BF suggested reaching out to Mathematics Director of Teaching (Stuart King) in the first instance and offered his support in the matter.
OUTCOME: Decision postponed until the required response has been received.
ACTION: BF to contact Mathematics Director of Teaching - Stuart King to seek clarification on whether or not the course MATH11176 - Statistical Programming can be made available to Informatics MSc Students in Data Science.
18112020_BoS Item 2 - New course proposal: Case Studies in AI Ethics (CSAI) - Nadin Kokciyan
Nadin Kokciyan (NK) presented her proposal on a new MSc level course, open to all Informatics students, aimed at tackling socio-ethical issues that arise in AI technologies.
The proposal was introduced to increase awareness around these topics with the intent to fill the gap within the existing curriculum. NK stressed the importance of facilitating an active discussion on the social and ethical issues arising in AI systems deployment. She noted a growing interest in offering lectures covering these topics was noted, based on the feedback received from Lecturers and Course Organisers in response to an email sent to teaching staff earlier this year. The PPLS-run course addressing similar issues (Ethics of Artificial Intelligence - PHIL10167) was quoted as a popular choice attracting a high level of interest, also amongst the Informatics students, as confirmed by the CO – Mark Sprevak.
The proposed course delivery: The proposal is to run the course in Semester 2 with the expected number of students of approximately 50-80 given no specific pre-requisites required.
The planned assessment: Exam(written): 80%, CW: 20%
Anticipated resource requirements: 2 case-study-based tutorials over the semester, with the expected 10-15 students and 1 tutor per each group; the number of the required markers dependable on the class size.
BF thanked NK for presenting her proposal and invited the Board members to comment on this.
There was a discussion around the target audience and potential prerequisites. Ian Murray (IM) asked whether basic Maths skills and CS or AI background would be set as course prerequisites. NK answered that the assumption was that the course would be for Informatics students, hence CS and AI would prove to be helpful. BF suggested that it could be worth being more specific on the target audience to ensure that the incoming MSc students have relevant AI background, given the fact they may be coming from other disciplines such as Physics or Mathematics. Sharon Goldwater (SG) proposed restricting the course to a certain degree, at least initially, to make it more manageable. It was noted that this could also facilitate the process of setting the course pre-requisites. Furthermore, it was suggested that a survey was conducted to gauge the level of interest and the prospective audience which may have a significant impact on how the course runs. Myrto Arapinis (MA) commented that she understood that this course was mainly designed with AI and Data Science students in mind, however, she felt that it would also be highly relevant to and much welcome by Cybersecurity, Privacy and Trust MSc students. MA also noted that there was another course tackling similar topics, seemingly run by the School of Law, which appeared to be a popular choice, yet one posing major enrolment difficulties due to being a capped course with the capacity for 20 students only. In light of this, AM asked whether the anticipated survey could be extended to students enroled on the aforementioned courses.
Queries were raised around the material and proposed course delivery. Bob Fisher questioned the assessed coursework and the depth of material. NK responded that although there would be no practical toolkit as such, the proposed content would include case studies based on “ethical framework” which is, effectively, a method of asking the right questions when tackling the problems encountered during the design and implementation process. Bjorn Franke (BF) queried the relatively low number of course tutorials and small groups' capacity (10-15 students per tutorial). NK clarified that lectures, based on presenting relevant case studies, would also involve an element of a discussion, hence a smaller number of standard tutorials would be required. She felt that 2 tutorials held in small groups would be sufficient, yet, this number could be reviewed and increased as required.
James Garforth (JG) commented that Informatics students don’t necessarily have enough background to engage with courses like this in the way that other students can (namely those enroled on Philosophy courses), which he felt can be built through active discussion. He also expressed his concern that the proposed course structure with only 2 tutorials and the large part of the material being delivered through lectures to a potentially large group of students may not give them enough opportunity to think through the course material and practice the skill. The approach suggested by JG received a cautious welcome from the Convenor. BF commented on the multifaceted issue of the resource-intensive course models, namely potential difficulties with funding and recruiting tutors with relevant background, emphasising the need to factor in the current economic situation within the School. It has been stressed that now that business case is to be taken into account more than possibly ever before, the level of teaching support should be carefully considered. JG counterpoint was that it might be difficult to establish what the current limits are unless an attempt is made to try and recruit enough support staff. BF commented that tutorials may not necessarily be the best for the newly proposed course.
A further suggestion was made that newly proposed courses should not be seen as static in this regard and as such should come back to BoS over time, with the first year being seen almost as a trial period. JG responded to the comment stating that his suggestion was intended to be feedback to NK rather than an objection to the presented course model. NK counterpoint was that as a starting point, 2 tutorials appear to be sufficient, however, there would be certain flexibility on this and numbers could be increased if necessary. She emphasised that her preference was to have a fewer number of tutorials delivered in smaller groups rather than more sessions offered to more crowded groups.
Bob Fisher (BF) asked how the implicit ethical concerns of students coming from various backgrounds would be addressed in the course. NK acknowledged the relevance of these and advised that these would be considered and incorporated when the course is fully developed. Further to this, NK noted that these may in fact be tackled as part of the current initiative of Decolonising the curriculum.
Michael Hermann (MH) commented on the potential benefit of the mixed audience and suggested introducing “either-or” condition in some respect. SG commented that although this approach could certainly be beneficial from the strictly academic perspective, it was likely to pose a risk of overcrowding, which affected a number of courses within the Informatics. She pointed out that similar courses offered across the University (including AI Ethics run by Philosophy) appear to be restricted (either due to prerequisites required or as a result of being capped), which directly translates into a very limited availability offered to Informatics Students. In light of this, the suggestion was made to restrict the proposed new courses (where relevant), at least in the first year, with a view to opening them up to a broader audience over time.
Further advice and feedback were shared via chat, where Oisin Mac Aodha’s, currently Tutoring for FDS, commented on the positive student response to the recently run tutorials - “Ethical discussion” where 2-3 case studies were provided to facilitate student active discussion. A suggestion was made that NK reached out to David Sterratt and Kobi Gal to discuss these.
Stuart Anderson (SA) commented on the wider subject. stating that in his view students have largely “synthesised” ethics over the last few years. He felt that the proposed model of delivering case studies appeared to be a good way of navigating through these issues and pointed out that understanding argumentation and critical thinking were vital, hence it could be worth embedding these into the course structure to help students build on their skillset in this area.
OUTCOME: Approved in principle. Conditional approval based on the suggested survey results.
ACTION: NK to arrange for a survey to gauge the prospective audience.
18112020_BoS Item 3 - Preliminary course proposal: Methods for Causal Inference - Ava Khamseh
Ava Khamseh (AK) presented her proposal on a new UG level course open to all Informatics students with previous experience of MLPR, focusing on the development and application of causal inference techniques as an important emergent area in AI and data science which is believed to have broad applications in social and biomedical sciences. The proposal was introduced with the intention to fill the gap in the curriculum as research is moving away from merely associational statements towards cause-effect statements.
The target audience: students with interests in Data Science, Machine Learning, applications of AI in Biomedicine and Bioinformatics.
The proposed course delivery and anticipated resource requirements: 2 Lectures a week (over semester 2) with students expected to work on the coursework in their own time, with office hours offered (2 hours per week)
Currently, a pilot version of the course is run with approximately 25-30 students attending informal lectures. ~70% of lecture slides have already been made available, together with corresponding recordings. Additionally, a website has been set up to provide further resources to students.
The planned assessment: Exam(written): 80%, CW: 20%
BF thanked AK for presenting her proposal and invited the Board members to share their views.
There was a discussion around the target audience, potential prerequisites and the course delivery. SG asked for some clarification on this, to which AK responded with a statement that these would be MLPR or similar. SG commented on this, emphasising that MLPR is one of the courses that students often struggle with, which could have a direct impact on the proposed course.
BF presented a comment from the Head of School (Jane Hillston) who advised that one of the Post-Doc Students reported finding it hard to keep up with the material, which sparked BF concern. AK responded that the feedback received so far from students who engaged in the pilot version activities was very positive and noted that it would be hard to comment on this without knowing the background of the student.
Bob Fisher asked where this course would be positioned given the fact that it fits somewhere between Maths and Informatics. AK responded that the Mathematical component in the course is not large and that students wouldn’t be expected to work on sophisticated models or perform complex estimations but rather on basic theorem, with a few variables, performing basic probabilities and making independent statements. The planned lectures and tutorials are to be based on real-world examples. During tutorials, students would be presented with a confounder, such as age or gender, and expected to go from associational statement through the relevant steps to the causal statement.
SG raised a question on how the full course delivery and its actual content may differ from the pilot version and what kind of exercises and support students would be given which could assist them in translating the theoretical content into practical work. It has been noted that PhD students tend to have a very different approach to lectures presented with no specific examples compared to UG and MSc students, and while the former is likely to engage in some individual research, the latter may be significantly less eager to do so. AK responded that the lecture is meant to be structured in a way that for every 4 slides of theory there is one specific example based on actual data and/or a case study to allow students the practice of forming statements. Furthermore, she clarified that tutorials are expected to follow the same delivery model with regular feedback provided. She also noted that similarly to statistics and MLPR-type course, a case study and relevant data set would be given, with questions broken down into smaller steps guiding students through the course material with the aim of encouraging analytical thinking. Furthermore, it has been pointed out that it is expected that this formula can be applied to varied fields such as Economics or Medicine and simplified, if necessary, by providing a fewer number of variables.
SG asked for further clarification on whether the quoted examples would be provided to students by the CO while presenting or given as exercises to work on. AK responded that both would apply. SG also queried two take-home exams noted in the proposal to which AK responded saying that this has since changed, and it should now read 80% exam, 20% coursework.
There was a chat discussion around the course size and suggestions have been made to restrict it in the first instance which BF was also supportive of, suggesting that further discussions on this should be held. Commenting on a number of positive responses received, he suggested agreement in principle, welcoming the full proposal from AK. In response to this, SG suggested that a separate discussion should take place beforehand so that the curriculum is agreed upon in the first instance. BF responded stating that although this was indeed the case, due to the current economic climate, all new course proposals are currently subject to additional discussions which should be considered as an entirely separate matter and as such should not concern BoS members when considering new courses and changes to the existing ones in general.
OUTCOME: Approved in principle, with a full course proposal welcomed.
ACTION: AK to submit a detailed proposal for to next BoS.
18112020_BoS Item 4 - Updates to Professional Issues - James Garforth and Stuart Anderson
James Garforth provided a context to the proposed changes stating that these have been triggered by UG Curriculum updates which are due to be implemented wef Semester 2, A/Y 2020/21.
JG advised that given the fact that much of the material currently taught in this course is due to be taught earlier to Software Engineering and Professional Practice (SEPP) students (in year 2 as opposed to year 3), the course itself cannot run in its current form and as such requires some updates. The proposed change was summarised as a shift in the prevalence of the topics and principles covered within the course. JG pointed out that the course currently follows ACM Code of Ethics and Professional Conduct, focussing on Section 2 which is “Professional Responsibilities” and the proposed change, is to focus on sections 1 and 3 instead; these are: “General Ethical Principals” and “Professional Leadership Principles”. JG concluded the proposal with a statement that the change felt as minimally invasive to the course delivery. It has been noted that JG is due to be the PI course CO in session 2021/22 since Stuart Anderson will be taking sabbatical leave.
A question was raised on the course's target audience and pre-requisites. Bob Fisher (BF) asked about the potential shift of emphasis within the course in the context of the material covered by SEPP, namely the potential increase in the number of students enrolling on the course with no previous experience of SEPP, who could, effectively, be missing relevant course material. JG explained that his understanding was that every student enrolling on the course would have already completed SEPP, with the possible exception of a few joint degrees. However, it has been noted that this would be unlikely to pose a major problem as the course is not building directly on SEPP. SG presented a comment on behalf Philip Wadler (PW) who sent his apologies for the meeting. PW asked for clarification on the subject of sexism and racism, which he felt were increasingly important, querying where in the combined course (SEPP/PI) would these issues (in particular, systematic racism) be dealt with. JG felt that these issues would be included in the course material and invited comments from Stuart Anderson (SA) on how broadly these were covered in the existing version of the course. SA confirmed that these were presented from different perspectives in PI and were also at the core of SEPP. It has been noted that professional responsibility tends to bring a certain focus to students in general, increasing their awareness of the relevant issues.
There was a discussion around the anticipated resourcing. BF asked JG to comment on the plans regarding this aspect. JG advised that all numbers were to be seen as flexible, yet, given the material, tutorials would be essential to the effective delivery of the course. He explained that teaching students to be able to discuss topics in a manner that may be familiar to Philosophy or Social Science Students but scarcely known to Informatics Students requires practice, which is best applied through the tutorials involving discussion. A reference was made to Oisin Mac Aodha’s comment in the chat on the positive response to the recently introduced “Ethical discussion” tutorials where, as noted by JG, students present on a rotating basis to the rest of the group. JG advised that this tutorial structure has also been adapted by Philosophy and suggested the implementation of a similar model for PI. It was noted that some level of pushback with regard to the proposed PI resourcing was anticipated.
BF raised his concern on the proposed additional tutoring resource, stressing the importance of factoring in budget and resourcing constraints given the fact that the School has recently been operating in a different financial environment which may continue to be the case in the nearest future. Additionally, the longstanding difficulties around recruiting tutors with the required background and skillset were noted. The JG counterpoint was that there was a plan to double the size of the tutorials, hence there would be no significant increase in tutoring resources required. BF responded noting that there was an ongoing discussion about a new model for allocation for teaching support. He advised that at this stage it has been unknown what the actual resourcing for each course this year might be as broader discussions on this matter are yet to be held. JG clarified that there would be some flexibility with regard to this, saying that the number of tutorials could be reduced, if necessary, to accommodate budget restrictions and other relevant factors, however, he felt that this would heavily correlate with how well the course material would effectively be delivered. A further point was raised on the potential impact of this on the accreditation.
SG advised having previously encouraged JG to increase the size of the tutorials and expressed her agreement on the statement that excessive reduction could have a detrimental effect on the course delivery. She raised her concern regarding marking as another potential resourcing issue, asking for clarification on what the current plan was in terms of marking arrangements, namely what type of coursework students would be expected to submit, what would they be getting the feedback on, and from whom and also what would be assessed, with the view of estimating how much marking effort would be effectively required. Anticipating that projecting the numbers at this stage could prove difficult, she asked that the broader plan details were shared so as BoS members could provide their comments and feedback based on their experiences.
JG explained that the projected marking would be based on regular short essays (approx. every 2 weeks) handed by students as part of the CW with the opportunity for other students to comment and feedback on those, which would give them an opportunity to practice their writing and build on their skillset. The final assessment idea would be to allow students to cherry-pick 2 of those essays to submit as final hand-in for the course. It was anticipated that this task would be most resource-intensive in terms of the marking effort involved. SG commented that she was supportive of the proposed idea in principle, however, she expressed her concern with how useful and welcomed peer feedback would be. Additionally, quoting feedback received from those who engaged in similar activities, SG pointed out that the arrangement may generate the same level of work for the CO that the standard marking would.
BF proposed that a decision on the course was postponed until the next BoS meeting. He commented that, based on the discussion and feedback, he felt that there was space for PI in the curriculum, however, given that PI is no longer strictly required due to accreditation, and therefore may now be run voluntarily, following a request received from Head of School (Jane Hillston), a business case should be made.
There was dissent as some Board members felt that the business case should not be relevant in this instance. SG asked for clarification as to why it could be necessary, given the fact that PI was an existing course which, in her understanding, merely required some level of change as it could not be run in its current form. She stressed that it would appear that a general feeling within the School was that students should be educated more on those issues, hence it would be very surprising for her to see this course being removed from the curriculum. BF responded that although he was generally supportive of keeping the course in the curriculum, this has been specifically requested by Jane Hillston. He also stressed that in light of the undergoing review of the Teaching Support provision and that given the recent change of the financial framework within the University, the course evaluation may be required, particularly in view of the substantial marking effort and costs involved given the course size (anticipated number of students: 220). SG counterpoint was that it could be worth considering in the first instance whether or not PI would in fact require more resourcing than any of the other courses that would be offered to students as an alternative. SA raised his concern about the suggested outcome. He emphasised the fact that PI has already been made a case and in his view, this was merely the question of which approach the School would be willing to take on this matter going forward, namely whether or not enhancing student awareness of the relevant issues by providing a reasonable training in professional ethics and code of contact would be considered as a complementary and valuable addition to the other, more technical-oriented courses within the curriculum. It has been suggested that JG and SA could help with producing costing shall this be helpful in the decision-making process.
The proposed costing projection was welcomed by BF, who explained that the term “business case” is in fact an official terminology used in the University regulations in the context of resourcing and costing issues, hence its application in this matter.
OUTCOME: Decision on the proposed updates postponed until a business case is made and considered by the School Executive.
ACTION: JG and SA to provide PI costing for the suggested changes.
ACTION: BF to seek Jane Hillston’s broader view on the matter and to share the outcome with JG, SA and the Board members in due course.
18112020_BoS Item 5 - Decolonising the Curriculum - Björn Franke (NB: Item be discussed more broadly January/February 2021 BoS Meeting)
BF provide a quick update on the initiative of Decolonising the Curriculum being a new initiative currently run within the School and the wider University. He advised that it has been decided that a Working Group will be set up with the aim of producing a set of guidelines on inclusion and equality. The Working Group Members will work in direct liaison with Course Organisers who, in future proceedings, will be required to make inclusion statements on new cases that come into the Board.
In one of the next BoS meetings, presumably in January/February 2021, a number of statements will be discussed which are expected to be provided by the Working Group.
Some comments were raised in relation to the initiative itself and possible alternative approaches to tackling relevant issues. BF advised that consideration of these matters was not within the BF and BoS remit, inviting Board Members to feed any comments into a parallel Working Group.
NK provided further insights on the matter as a current member of the Working Group. She briefly explained that the idea behind holding the workshops was to encourage reflection on the curriculum and enable the University teaching staff to actively engage in a discussion and share their ideas during workshops.
OUTCOME/ACTION: N/A - Item be discussed more broadly January/February 2021 BoS Meeting
BF thanked everyone for the active discussion and concluded the meeting.