The results underscore another growing problem in AI, too: the sheer intensity of resources now required to produce paper-worthy results has made it increasingly challenging for people working in academia to continue contributing to research. “This trend toward training huge models on tons of data is not feasible for academics…because we don’t have the computational resources. So there’s an issue of equitable access between researchers in academia versus researchers in industry.”
The article focuses on the scale of the financial and environmental cost of training natural language processing (NLP) models, comparing the carbon emissions of various AI models to those of a car throughout its lifetime. To be honest, this isn’t something I’ve given much thought to but to see it visually really drives the point home.
As much as this is a cause for concern, I’m less worried about this in the long term for the following reason. As the author’s in the article stake, the code and models for AI and NLP are currently really inefficient; they don’t need to be neat and compute is relatively easy to come by (if you’re Google and Facebook). I think that the models will get more efficient, as is evident by the fact that new computer vision algorithms can get to the same outcomes with datasets that are orders of magnitude smaller than was previously possible.
For me though, the quote that I’ve pulled from the article to start this post is more compelling. If the costs of modeling NLP are so high, it seems likely that companies like Google, Facebook and Amazon will be the only ones who can do the high end research necessary to drive the field forward. Academics at universities have an incentive to create more efficient models, which they publish and which then allow companies to take advantage of those new models while at the same time, having access to much more computational resources.
From where I’m standing this makes it seem that private companies will always be at the forefront of AI development, which makes me less optimistic than if it were driven by academics. Maybe I’m just being naive (and probably also biased) but this seems less than ideal.
Two weeks ago I presented some of my thoughts on the implications of AI and machine learning in clinical practice and health professions education at the 2018 SAAHE conference. Here are the slides I used (20 slides for 20 seconds each) with a very brief description of each slide. This presentation is based on a paper I submitted to OpenPhysio, called: “Artificial intelligence in clinical practice: Implications for physiotherapy education“.
Students working in small groups interact in a variety of ways and the teacher has an important role to play
Barriers, more often perceived than real, may impede the adoption of small group teaching
Small group learning is the learning that takes place when students work together usually in groups of 10 or less
What matters is that the group shows three characteristics.
Active participation: A key feature of small group work is that interaction should take place among all present.
A specific task: There should be a clearly defined task and objectives, and they should be understood by all members of the group.
Reflection: In small group learning, it is important to learn from an experience and to modify behaviour accordingly. Deep learning is a key feature of small group work: reflection is a key feature of deep learning.
It provides students with experience of working in a group, it helps them to acquire group skills. These include the ability to communicate effectively, the prioritising of tasks, the management of time and the exercise of interpersonal skills.
The introduction of small group work into a curriculum is frequently resisted. Various arguments are used
“Students do not like small group work”
Initial dissatisfaction with small group work may be expected and is natural. Barriers may be:
“Staff do not know how to teach in small groups”
: Teachers may lack the skills necessary for running small group sessions. This may be seen as more of a problem by course or curriculum organisers than by teachers themselves. A staff development programme, however, is important.
: Staff shortage may be a real issue – or may be a misconception. The small group method adopted and the timetabling both have a significant impact. Identify all the staff who could be available for small group teaching.
“We do not have enough teachers for small group work”
“There are too few rooms”: Space is frequently a contentious issue. Creativity is usually the best solution. Do not be afraid to experiment – students are resilient.
“It is a waste of time – students do not learn anything”: It may take longer to cover a topic in small group work than in lectures. However, what really matters is if, and what, students learn and not just what is taught.
The following checklists have been generated for running small group work.
Consider the objectives of the session: Consider carefully the objectives of the session or course you are running and whether other teaching methods, eg lectures, independent learning, may be more appropriate
Determine your available physical resources: Accommodation availability may be a limiting factor. Small groups require suitable accommodation, which allows chairs to be set out in a circle to maximise interaction among the students, and space for appropriate audio-visual materials.
Determine the manpower availability: Small group teaching requires the participation of a larger number of teaching staff. Consider carefully how many teachers are available and their expertise as small group facilitators.
Does a facilitator have to be an expert? For small groups to function most effectively, the facilitator should have expertise in the content area and in small group facilitation.
Can different facilitators be used with the one group? Continuity of the facilitator is desirable. The facilitator can be considered as a member of the group and changing the composition of the group may disrupt the group dynamics.
Can a teacher facilitate several groups? A floating facilitator can be assigned responsibility for maintaining the task and function of more than one group. Success varies, frequently depending on the dynamics of the groups and the skill of the facilitator.
Can a student facilitate his or her own group? A student can successfully facilitate a peer group. This depends on the group’s dynamics, on the student’s ability and on the briefing given to the student.
Determine the group membership: Groups may be self selected, strategically determined, randomised, or selected alphabetically. It may be appropriate to change the membership of a group after a designated period of time, eg semester or year. Groups will work less productively if constantly changed, as they are less likely to reach a productive stage.
Ensure that the staff is prepared for the session: Staff must be given detailed briefings for the small group work. They should have prior opportunities to see small group work in action and to attend staff development sessions.
Select the most appropriate small group method: The educational objectives should determine the choice of small group method.
Develop stimulus material: Stimulus material might include problem-based scenarios, video clips, a set of questions, key articles for exploration, a real or simulated patient. The stimulus material required will be determined by the small group method adopted.
Inform students about the role of the small group work: Students should be informed why small group work is being adopted and how it relates to other activities and to the learning outcomes for the course.
During the Small Group Activity
Allow adequate introductions – use ice-breakers if necessary
Ensure that the students understand what to do, why they are doing it and how they should achieve it
Facilitate learning: The facilitator’s objective is to help the student become more self-reliant and independent by establishing a climate that is open, trustful and supportive. An educational facilitator’s role has two distinct areas: maintaining the functioning of the group and ensuring the task is completed. Facilitators must understand the changing dynamics of a group through four stages: forming, storming, norming, and performing
Debrief the group on the activity: Debriefing summarises or clarifies what has been learnt and may take as long as the activity itself. During debriefing, constructive feedback may be given
After the Small Group Activity
Evaluate the success of the session: There are two aspects – achievement and quality. Have the objectives been achieved? Was the educational experience of a high standard? Students may be asked to complete an evaluation questionnaire. Alternatively, the session can be peer reviewed. Teachers should consciously self-evaluate the session.
Reflect on the experience: Evaluation, formal or informal, is pointless if no change in practice results.
Small group methods have a valuable role to play in undergraduate medical curricula. The student-centred focus and active participation enhance the likelihood of deep rather than surface learning. Decisions to be made are how much time to schedule for small group work, and what kind of small group work to adopt. If small groups are used, they must be valued and must not be seen as isolated from other aspects of the curriculum – or from the culture of the medical school.
Staff development is important. Success of small group learning depends on good planning, effective facilitation and a movement away from teacher-centred learning. For a real and sustained shift in medical education, students must be encouraged to learn rather than merely catching the output of teachers.
I’ve been trying to think how to use technology to enhance both my teaching and my students’ learning and it’s proving more difficult than I’d initially thought. I like to think that laptops and internet access in every classroom give students real-time access to related content while they engage in meaningful discussion, but this will never happen. Their Facebook profile and IM conversations are far more interesting than the “Pathology of stroke” or “Justice in access to healthcare”. And that makes sense in a bizarre kind of way. Even while they (or their parents) pay vast sums in tuition fees to have the privilege of attending university, most students (in my very limited experience) see studying as inherently boring.
Some studies in American classrooms have all but proven that the distraction of the Internet in class is too strong for students to ignore and that most of the lesson is spent checking email, catching up with friends and even shopping. Now, after that initial foray into “embracing” technology”, it seems as if there’s a move towards banning laptops altogether.
This is the kind of about-turn I’d like to avoid. E-learning, while I have no doubt will be a revolution in education, is not the idea that technology for it’s own sake is the way forward. Just because it’s possible to have Internet access in class, does it mean that we should? Rather, teachers must take an approach whereby technology is used in a way that enhances it’s advantages, while minimising the disadvantages. Just because I put the course reader online doesn’t make it “e-learning”, and neither does having a student blog. The technology in itself doesn’t enhance learning in any way, but how you use it can have powerful implications.
I’ve been toying with the idea of using a wiki to manage a course, whereby any change to either the course content, test schedule or mark availability can by syndicated through RSS to all the students in the class. Students will have to, as a course requirement, both add to and edit course content (obviously moderated), which can also then be tracked. I think that this may be one way to encourage them to actively engage with the content, as well as introduce concepts like peer review, referencing and drafting, which may also improve their reading and writing skills (another huge problem). The point though, will be to make the learning outcomes apparent from the beginning, so that students know what’s expected of them. Merely creating a wiki and telling students to “Go forth and create content” isn’t enough.
I think that technology will fundamentally change the way we teach and how students learn, but not just by throwing technology at the problem. The trick is to figure out how to use technology to facilitate deep learning by getting students to actively engage with the content. A bad teacher will continue to teach badly, no matter how much “technology” they use.