CIRCL perspectives offer a window into the different worlds of various stakeholders in the cyberlearning community — what drives their work, what they need to be successful, and what they think the community should be doing. Share your perspective.
At Apprendis, she and her colleagues created the Inquiry Intelligent Tutoring System (Inq-ITS, funded by the US Dept of Education and the NSF). Inq-ITS is a web-based learning environment that allows students to “show what they know” about inquiry. Students use simulations to perform NGSS-aligned inquiry tasks, and while they do, their work products and processes are assessed in real time using patented algorithms.
A companion tool to Inq-ITS, Inq-Blotter (funded by the NSF and the US Dept of Education), sends real-time information about students’ competencies on the NGSS inquiry practices to the phones or tablets of teachers. These real-time alerts allow teachers to know how each student is doing and whether the student or class understands a specific practice or not. With this information, teachers can change whole class instruction on the fly or help individual students who need it if only a few students are struggling.
How did your research lead into starting a company?
It’s kind of a complicated story to tell. We developed the initial microworlds for physical science on an NSF DRK12 grant (in 2007) and then we were funded by the US Department of Ed for he Life and Earth Science microworlds. NSF and the Department of Education then funded parallel projects, so we could develop a pedagogical agent to help students conduct inquiry for each of the content areas. Then we (Apprendis) won two SBIR grants (PI: Mike Sao Pedro) to develop Blotter, our teacher dashboard, so that student data from Inq-ITS can inform instructional practices. The NSF Cyberlearning program also funded research on Blotter to better understand teacher-student discourse that occurs when feedback is provided via Inq-Blotter. The research grants and the work development work on Inq-Blotter are like pieces of a puzzle that fit together.
How did you know you had a commercializable product?
You have to think about the ecosystem, what your product can do, where you think the technology will be like in future classrooms, and what policies are in place. Then it’s sort of like having a crystal ball. For example, I was partly inspired by the “quantified self” movement; it’s not hard to imagine that if you had technology tracking what a person does, where their eyes are, their effort, their knowledge, their skill acquisition, etc — and then analyze and react in real-time to all those things, that that would be extremely powerful. You also need to be aware of the state and federal policies around individualized and personalized learning. Generally, it’s helpful to broaden your lens, step back, and look at the ecosystem of what’s out there and what could happen. Think about what is in schools now and what they lack in order to support real time, personalized learning, then imagine a system undergirded by technology (with an eye towards how bandwidth, etc., will change in the near future), and then design towards that.
What’s different about your technology?
We have developed proprietary (patented) algorithms that use machine learning to automatically grade students’ inquiry work. Our research has shown that it agrees with human scorers’ grading up 95% of the time. Our algorithms also do a better job than general Natural Language Processing algorithms, whose match to human scorers is typically about 80%. We get such high accuracy because we use a theoretical lens to guide the development of our scoring. Our theory is based on the inquiry sub-skills we have identified underlying the respective practices. NGSS and inquiry can be time intensive for the teachers in the classroom, so these kinds of computational technologies can make NGSS realizable in terms of assessment and personalized help and they allow for implementation at scale. You have to give up a small amount of information about context, but you gain a lot in real time reactivity, personalization, and of course, scale.
What’s it like to be a professor and start a company?
If you’re interested in working about 20 hours out of 24 — if you have that bandwidth — it’s about that intense. The other key factor is that you have to have really good, interdisciplinary teams. You’ve got to have your eye on the vision research-wise, your eye on the product development vision, your nose to the grindstone, and also know when you have to pivot from a business perspective. That’s the interesting thing about business and technology development, you have to know when there’s something new worth paying attention to, and you can only do that if you’re attuned to everything being developed.
If I walked into a learning environment that was using your innovation, what would be different?
I have two answers to that question that are important. There are about a million simulations out there, but what’s different about our simulations is that they are undergirded with patented algorithms that react in real-time to the user and give very fine-grained performance metrics to the teacher in real time.
Our technology uses authentic measures to understand if a student understands how to run, for example, “controlled trials”, as well as all other NGSS practices. In the case of checking whether the student is running controlled trials, our algorithms take into consideration all the trials the student run, regardless of whether they’re sequential or not and determines whether or not the student knows how to run controlled trials; other techniques that do not use machine learning do not allow for this flexibility. What is great about that is that for every inquiry practice, the teacher is getting real-time feedback on each student and can stop the class in real-time, or talk to a particular student to help them.
The second answer is to tell you about a teacher in Arizona who is using our tools, and finds them truly transformative. She’s actually in treatment right now, sadly, for cancer, and she’s using our technology to monitor her kids while a substitute teacher is in the classroom and she is in chemo. She’s using our tools to write her kids in real-time, saying things like, “hey Johnny you know you’re doing this”; “Hey Billy you’re doing that”. She can support these students remotely.
The time for real time feedback is in the moment, while the student is learning. It doesn’t make sense to allow students to go through all the inquiry phases and then a week or two from then get back to them. It’s so important that students get the feedback in real-time and the teacher can stop the class in real-time. It affects a teacher’s instruction in a way that supports inquiry learning as it needs to be supported, i.e., in the moment. It’s like apprenticeship learning where you provide situated experiences with feedback in real time, with the apprentice working side by side with the mentor; that’s the big benefit we see from Inq-ITS and Inq-Blotter.
What do you think the cyberlearning community should be doing?
It’s both encouraging and discouraging to go to a cyberlearning meeting and see such great tools, and then wonder, “Why can’t this can’t be in the hands of more people?”
This gets back to the need to commercialize. As researchers we often say, “Why commercialize? Aren’t we working on the government’s funding? Shouldn’t our work be free and freely available?” But that’s why there’s the Bayh-Dole Act that allows the ownership of inventions made with federal funding. Our responsibility should be to get our good work out there in the hands of teachers and students. Cyberlearning community members are doing really good and innovative work. They should commercialize. We don’t want products out there that aren’t based on research, and if we don’t do commercialize good, research-based products for teachers and students, who will? That’s why I work so hard. I want the good work of my team to be out there for all students.
Also as academics, we like the Platonic ideal to exist before we put it out there, and that’s why some won’t work toward commercialization. Many want solutions that are 100% perfect. But every single year we have students graduating from high school hating math, or hating science, and we’re basically employing people offshore for our STEM jobs. I think that we’re really at this transformative time where you have a lot of groups working at the intersection of computer science and cognitive science/learning sciences and we have amazing cutting-edge stuff that tells us a lot about learning. We need to leverage these projects to put better products into classrooms. As Voltaire said, “Don’t let the perfect be the enemy of the good!”