Subscribe
Share
Share
Embed
The Transform your Teaching podcast is a service of the Center for Teaching and Learning at Cedarville University in Cedarville, Ohio. Join Dr. Rob McDole and Dr. Jared Pyles as they seek to inspire higher education faculty to adopt innovative teaching and learning practices.
This is the Transform Your Teaching Podcast. The Transform Your Teaching Podcast is a service of the Center for Teaching and Learning at Cedarville University in Cedarville, Ohio.
Ryan:Hello, welcome to this episode of Transform Your Teaching. In today's episode, Dr. Rob McDole and Dr.
Ryan:Jared Pyles chat with one of our favorite people in the whole world, Dr. Haisong Ye. Dr. Ye is an instructional designer here in the Center for Teaching and Learning, and we are interviewing him as a needle mover. Thanks for joining us.
Rob:Well, Jared, we're here on a one off with one of our compatriots and colleagues, as part of our Needle Mover series where we have some folks come in who we feel like have shown a demonstrated way to move the needle in terms of teaching and learning. Mhmm. And today, we have doctor Haisong Ye.
Jared:Whom I believe I've met before.
Rob:Have you have you two met
Jared:I think we've met
Haisong Ye:Every day?
Jared:Every day for the last eight years, seven or eight years, something like that. So
Haisong Ye:Yeah. Thank you for inviting me.
Jared:We just wanna have you on because we, at least I highly respect you and think the world of you, and you have pushed me in ways that, have been very challenging, especially when you just look at me and go, why? When I have an idea for something or, you know, have a problem, he's like, just fix it. And then he'll tell me how to fix it, which is especially with with us using ChatGPT and stuff. It's been a honor and a privilege to, work with you and learn from you. So
Rob:Well, doctor Yeh, one of the things that you're... You've brought to the team is is just a real strong technical sense of what we're able to do here. And I would say you've made it very possible for us to continue to do what we're doing, even this podcast, because the amount of work that we've had over the years has grown significantly. It hasn't lessened. And it seems like the more we find ways to automate, the more things we find.
Rob:Yeah. And you've been fairly instrumental in that for us. We just wanna maybe peek into, maybe give our listeners an idea of some of the things that you've been into. So you teach database as an adjunct here for the institution. Yep.
Rob:Which kind of signals where your technical abilities are at. And why don't you talk to us a little bit about how you're using AI, like maybe some of your favorite projects and... Yeah, let's just start with that.
Haisong Ye:So I think one of the favorite projects would be the course auto build in Canvas. Okay. This is my favorite one because before that, we always need to... We build the course alignment grid first Right. At CTLP.
Haisong Ye:So we have all the course objectives listed, and we make assessments aligned to the objectives, and then we put those content, the resources to fulfill the assessment and align to the objectives again. And we have the completed course alignment grid. And then we need to manually click by click, create pages, create assessments, create modules, move things around, make some mistakes, mistakes. Right?
Rob:Check your mistakes. Yeah.
Haisong Ye:Then once once we had... No. That's before before Cedarville has OpenAI subscription. We had the OpenAI API stuff. Yeah.
Haisong Ye:We have the API platform, and then we can use Python to call OpenAI API directly to do things. At that time, we used Python to, like, to create those Python scripts to call the OpenAI API to read the course alignment grid, figure out how many modules there, what are the objectives, what are the assessments and the resources, categorize them into, like, different types of Canvas objects or elements, like pages, quizzes, discussions, assignments, rubrics. Once we had that set up, we can use Canvas API again to automatically create those objectives... Objects or elements in Canvas Mhmm. Automatically.
Haisong Ye:And then based on the course alignment grid, we set up the modules based on that. That frees us a lot of time to manually go through the course alignment grid and to create those elements. Will say
Rob:One at a time.
Haisong Ye:One at a time. Yeah. I will say based on my work, course build, usually, it will take at least two or three days. Mhmm. For some back, back and forth, check, copy and paste.
Haisong Ye:Now I think two to three hours Yeah. Yeah, I can build a course.
Jared:And even then with the Python, when we did it that way, it would just create the shell. We still have to go in and copy and paste all the assignment
Haisong Ye:Mhmm.
Jared:Stuff then. But since we started using and you have this auto builder, now that's like upgraded now in the Canvas skill. Yeah. Like, we can do a lot more than that.
Rob:Mhmm. Which
Jared:I... Which I've done myself. Like, it'll... You can start with the same thing where you... I attach the course alignment grid to our skill.
Jared:It will build it. Mhmm. Then after that, I'm like, okay. Verify all this stuff and then go in and I can then bring in the whatever assignment descriptions I have in Word docs and say, here's the assignment descriptions. Now populate what you just created.
Jared:Mhmm. And it does it. So, like, I I don't think our listeners know how... Well, maybe they do. How monotonous it is sometimes to build courses and keep clicking.
Jared:I used to
Rob:Clicky. Clicky. Clicky.
Jared:In order to make it fun for me, I used to speed run. I used to, like, set a timer to see how long it would take me to build assignment descriptions by hand and pages, and I tried to beat my time each time because I was so bored with the process. And I had to go back and authenticate and make sure that everything was accurate as as a speedrunner should. But but now it's like literally throw it in there and it will build it for you. There are some quirks, like you and I even as IDs, we have different styles that we like to Yeah.
Jared:Customize in our builds. So it just takes me... Usually, I have to just... I'll pull up an exemplar of the way it should look, and then it will build off of that exemplar.
Rob:Mhmm. Mhmm. Yeah.
Jared:But again, the amount of time we're saving because of the work that you've done is tremendous. And we've talked about this before, how it can free up time to do Right. The things that are important... More important to us than... Not saying these items aren't important, but there are more things that the human to human contact and relationship, we can now do more of that as
Rob:a result of this work. Well, we can definitely spend more time actually dealing with the important things, which isn't necessarily. I'm not saying that putting it in Canvas isn't important. It most certainly is. It's gotta be there.
Rob:Yeah. But the really the important thing is the interaction with the information. Mhmm. And and then... So there's having an evaluation is is it accurate just baseline, Which these tools have...
Rob:And we've been working with them and and doctor Ye has been working with them and he's... You know, we've continued to get better at directing these LLMs. The LLMs have gotten better, but our scripting has also gotten better and what we expect out of it and what we know it can and can't do. And with that has been a shortening of the iterative design processes, which Yep. I mean, essentially is what we adhere to here.
Rob:It's like a continual improvement design. So we know that the first version that walks out the door is never going to be the best version, and we're okay with that. But it does need to meet a certain level of standards, which it does. And then comes the important part is that as you try and you test it and you see, oh, okay. No.
Rob:This isn't gonna work. And what his tools have helped us do is to move quicker Mhmm. Down that pathway and to shorten the time it takes to go from prototype to viable product. And not not just how that'll do, but a high quality product, I think is exponential. Just like you said, like, go from three days to two, maybe three hours.
Rob:I know for myself personally, I just built two courses that I'm running and I use the same tools. You know, and I was able to make wholesale changes where I was sitting there thinking, if I made these editorial changes and I sent them to to to Jared, who served as the ID for my courses, and I said, hey, what do you think about these? And I asked him to do, you know, the updates on the course, he would rightfully lose his mind, the number of changes that I've made. Now, genuinely, you would just like...
Jared:No. I would just say no.
Rob:I'm not do Or you'd like, oh, we'll get a student worker to do that.
Jared:That's actually more accurate.
Rob:That would be part two. But, you know, now we can move a lot faster. And because the the number of courses that we're developing, you know, has has definitely increased. And the demand and the quality that we want out of them has not. So I think it...
Rob:It's well stated to say that the work that you've done has helped us move move along in a way.
Jared:Move the needle.
Rob:Most certainly, move the needle. Mhmm. Yeah. So what else? Is there anything else that that you've been doing that you think, yeah, I'm I'm enjoying this?
Haisong Ye:Another one would be the painful, maybe, course resets process.
Jared:Yeah.
Rob:Our dear project manager, Ryan Lyme, would probably agree with you on
Jared:that one. I'm also realizing as we talk about how quickly we can build courses now that our project manager, comma, Ryan, is going to go, why aren't these done then? You have all the stuff. Why isn't it being done?
Rob:We're waiting on someone.
Jared:Yeah. Waiting on me. Walk us through the course resets because that... Yeah, I had to do that during COVID.
Rob:And most people... Yeah. Most people don't know this exists because I think as an... As a team, you all have made this invisible for the most part to those who use these systems. They don't even think about it.
Rob:It's it's kinda like the road. You get in your car and you drive on the road, but nobody ever thinks about the road until they hit a pothole. Right? And I would say, to us, the course resets is a big deal. So, yeah, talk about that.
Haisong Ye:So the course resets is every semester for all the online course sections, both undergrad and grad course... Online course sections, we will have a... We have a blueprint master course for each of those online courses. And then every semester, we need to prepare a... Prepare live sections for next semester.
Rob:Mhmm.
Haisong Ye:Like, in spring, we'll prepare for summer.
Rob:Yep. Summer, we'll for fall.
Haisong Ye:Yep. So, usually, that will... Especially spring and fall, we will have, like, how many courses? Two hundred two hundred sections we need to copy from blueprint master template to the live sections. That includes adjust the dates from spring to summer, summer to fall, or fall to spring, like shifting the dates and make edits that requested by the course coordinator Mhmm.
Haisong Ye:To improve the course. So all these things adding together, we usually have one CTLP staff and two student workers to work on this for, like, two months.
Jared:Two months. Man. Because you have to go in and manually change every single date.
Rob:Yeah. It's a... It was a... It's a very, very, very long Yeah. Process.
Rob:You talk about clicking fatigue. Yeah. Mhmm. Because you gotta go through each one of the assignments and make sure according to whatever the dates are and the breaks are. Mhmm.
Rob:And make sure that that all sits properly.
Haisong Ye:Yeah. So although Canvas has some, like, batch edits functions, but it still take a lot of time. Right. And very easy to make mistakes. Mhmm.
Jared:Sure.
Haisong Ye:Yeah.
Jared:Yeah.
Haisong Ye:So we started thinking about it, how to... We have a whole workflow, how to process that course resets. I think Lauren and Ryan create a, like, five or six pages of workflow. Step one, do what? Step two, do what?
Haisong Ye:And then with AI, we go through the steps and identify those steps that can be or could be automated by script. And we pull this out, and then I think we did some testing over two semesters, and then created a web app that can read course schedule from register's office. Just pick those online courses in. Before Ryan would go to registered office, go to each department, find which section is online, copy that, and put it into a Google Sheet. Before it's Google Sheet, then later it's a Microsoft list.
Haisong Ye:That that alone will take, like, a couple weeks. Yeah. So that's that's the first step. We automatically grab the courses from register office, only online sections. Later, we added CSP and Christmas school partnership for listeners.
Haisong Ye:Yeah. Yeah. Yeah. And ORP, OPR, online program only courses. All those courses were categorized as online sections.
Haisong Ye:Mhmm. So we handle this all. And then once we have that, we will work on the master course. So we automatically, based on the course code link, find the master course in Canvas. That's why we had a naming convention Yeah.
Haisong Ye:Discussion. How how do we name our blueprint courses? Because we have a lot of archived eight week. Right. Like, mast eight, mast 12, mast 14.
Haisong Ye:Yep.
Rob:And a lot of different Versions. Versions. Yes.
Haisong Ye:Yeah. So we fixed that, and then we were able to link live section with brute... Blueprint courses based on based on name. And then we keep using the script I built before, like, for the course build... For the course builder and reuse those scripts to adjust the dates and then update the syllabus because they...
Haisong Ye:We have dates in syllabus. Mhmm. And once in a while, we may need to update the syllabus as well because we have new contents for syllabus. Textbooks. New textbooks, even new policies.
Haisong Ye:Yeah. And then we, again, use Canvas API to link the blueprint to link the blueprint with those live sections and push the content to the live section. And when the times come, we will automatically publish the live section for Wow. Preview week or something. Mhmm.
Haisong Ye:Mhmm. But it's it's it's working. We have used it at least for two or three semesters... Two or three... Two rounds right now, but we still feel...
Haisong Ye:Found a lot of
Rob:Opportunities for improvement.
Haisong Ye:For improvement. Yeah. Yeah. Yeah. Because our workflow is build built based on how we do it manually.
Jared:Right.
Haisong Ye:Right? Because our mind... Mindset is still based on how we would do it manually. But I think with cam... With codec skill right now, we can rethink about it.
Haisong Ye:Even from the workflow itself, we can change it. So that's our next step.
Rob:Well, and I think for our listeners, the thing you gotta understand is, like, the time to do this. Like you said, it was, like, two months for all the people involved just to get those sections out the door, and we're talking about hundreds of sections
Jared:every semester. Over 300 different sections for this fall, and it saved at least a 100 man hours. Yeah. That's crazy.
Rob:It is. And and then you think about what that saves us.
Jared:More time to make coffee.
Rob:More time. Well, more time...
Jared:You bought an espresso machine. And I think it's indirectly related to the work that doctor Ye has done.
Rob:It's a cheap investment.
Jared:A cheap investment. The amount of hours we now have available to us to make coffee is tremendous.
Rob:If only that were true.
Haisong Ye:It was not it was not the very... It was not perfect right now. You still have Sure. Complaints from students and professors, but I believe it's less than before. Yeah.
Jared:Yeah. For sure.
Rob:Mhmm. Yeah. When you actually think about, and and this is something a lot of people don't think about, like error costs. So when we get things right to, know, a better degree, that means we've got less on the back end to actually have to do to correct errors. And and every system, and I know this may be hard for some of our listeners to believe, but even here at Cedarville, we have errors.
Rob:The question is what do you do with those? Right? And you have to correct them. So the more errors you have, the more correction you have to do, which is more time. And if you're taking up time correcting errors, and that means you're not...
Rob:I mean, for us anyway, it's like we do one cycle, we have to turn around and do it again. Right? Because we've got eight week courses, so we've got subterms. So it's not just fall, spring, summer, it's fall a, fall b, spring a, spring b, summer a, summer b. So we, you know, we have what, six subterms, if you will, that have new starts.
Rob:And so those those sections have to be adjusted. We have to take in account any changes that needed to take place, and that's a lot of moving parts and a lot of moving pieces.
Haisong Ye:And we'll have a lot of exceptions.
Rob:There will always be exceptions.
Jared:Someone always has to be special.
Rob:And so we have to accommodate that. And and your work along with Ryan and his team, you know, have taken what was... What seemed overwhelming and probably gave gave some nightmares at times much more manageable. And so I would call that a needle mover.
Jared:Yeah, for sure. I think if we if we were to unleash what we do here at in CTLP to the general populace, because we... You and I have even tried to present what it is that we've done. And so far it's been like, yeah, we don't we don't wanna see that. But I I I don't think it's because it's boring or it's like necessarily not good because it's so good that it could upend what a lot of practices have been forever.
Jared:Like, I feel like we're at a point where we're at a bleeding edge for this Mhmm. Through doctor Ye's work. And now if we... Other people hear about it, it's like, oh, oh my gosh. You know?
Jared:So
Rob:Well, and I think it does for some some of those centers for teaching and learning and those those units out there that that deal with this stuff day in and day out. They'll understand if they're listening, you know, they'll understand what some of the stuff that we're talking about. But I would say k through 12, you know, understand it to to a degree as well because every year, there's been more technological demand on them, more layers of technology that have been put on them, more things that they're they're needing to layer into the day to day. And if you don't have automation or if you're not tackling some of these things with automation, then you're missing out and you're missing in moving that needle to be able to get back to actually asking fundamental questions about education and making sure that the tail is not wagging the dog. And I would say that's what our efforts have been here.
Rob:We do want to bring technology in.
Haisong Ye:Mhmm.
Rob:But we want it to serve us, not us serving it. And I think that's what some of this work has helped us do. And and if any of our listeners want to, you know, connect with doctor Yeh or connect with with us here at CTLP, feel free to reach out.
Jared:So, Doctor. Yate, we're very thankful for you, and, I'm personally thankful for you, like I said before,
Rob:I am as well.
Jared:Thank you. Thank you for, pushing us. I know I can speak for Jason as well when I say thank you for pushing us as IDs and as members of CTLP and the work you've been doing has been tremendous and it's a challenge to the rest of us.
Haisong Ye:Yeah. Thank you.
Rob:Thank you.
Haisong Ye:Yeah. You guys are very helpful as well, so we help each other.
Ryan:Thanks for listening to this episode of Transform Your Teaching. If you have any questions or comments about our conversation with doctor Yeh, feel free to reach out to us at CTLPodcastcedarville dot edu. You can also connect with us on LinkedIn. And finally, don't forget to check out our blog at cedarville.edu forward /focusblog. Thanks for listening.