Streamline Learning Curves in the Workplace with Machine Learning – USC Viterbi
Work smarter, not harder
Starting a new job can be difficult. Learning the ropes involves more than just mastering information and understanding the tasks you are expected to perform. Often there are role or company specific tips and tricks.
These useful hacks are more difficult to learn because they are not always self-explanatory. Instead, they are often found in the unique experiences and knowledge that only someone who has worked before has, which can be lost when transitioning from an old to a new employee.
After a few weeks or months at a new job, you gather enough information to be successful, but you may need to create your own methods for completing tasks or even consult with the person who held the position before you.
What if you could learn everything you need to know faster, without struggling with unnecessary tutorials or bothering a former employee or supervisor?
The KNIC (Knowledge Needed in Context) program at USC Viterbi’s Institute for Information Science (ISI) is developing a way to extract this valuable information from the former employee’s head and put it into a form that the new employee can use.
The Secret Ingredient – AI Helpers
The KNIC program is part of the Knowledge Management at Speed and Scale (KMASS) project that the US government Creation of the DARPA (Defense Advanced Research Projects Agency). (1)
Jay Pujara, director of the Center on Knowledge Graphs at ISI, said the idea behind KMASS was to build an AI that can “interface and watch someone do their job, learn from it like a apprentice and then store this knowledge so that it can then be provided to the new person and help them learn the job.
Thinking back to human history, implementing this “old way of learning” makes sense – following someone who’s mastered the job makes it easier to get over the learning curve, because you’re acquiring his own proven methods.
In this case, the skilled employer comes in the form of an AI “buddy” who is able to assess what the novice is doing and what they need help with, without boring them or hampering their productivity. .
“The basic research vision that we have is to build this companion that dialogues with you, either a producer of knowledge, performing a task and it observes you and tries to learn from what you have done, either as a consumer of knowledge and he observes you and he tries to give you some helpful suggestions,” Pujara explained.
Take typing a line of code, for example. This AI companion would be able to identify in real time when the person is making a mistake and suggest a solution or provide advice on where they could go for help.
It’s all in the knowledge
The KNIC is divided into three distinct elements: knowledge storage and organization, knowledge capture, and knowledge dissemination.
Knowledge store describes the process of compiling basic knowledge about a topic, drawing from laboratory and academic literature, the Internet, and other sources to establish context.
Knowledge capture refers to the AI companion’s interactions with a knowledge producer, learning what it is doing and determining what questions to ask and when to ask them during the task demonstration without redundancy and without irritating the person using the system.
The third area, knowledge dissemination, involves communicating and providing information to the knowledge consumer that would be useful in the context of what they are trying to accomplish and based on what they already know.
Elisabeth Boschée, Associate Director of the AI Division at ISI and Director of ISI Bostonsaid the team began their research by experimenting internally.
“We first had a master’s student at USC go through the documentation and try to create a model,” Boschee explained. “It didn’t pass, which was good in this case because it shows that the documentation produced by our own labs is flawed or aging badly.”
This “guinea pig” experience allowed the team to learn what type of information would have been useful for the system to provide to the student when she found herself stuck on a step.
But would a new employee really learn to complete tasks on their own if the bot was there to give them answers every step of the way? What about the classic “learning by doing” method?
Boschee claims that the KNIC program would actually facilitate learning by doing, as the technology would not perform tasks for the user, but would help them quickly find the knowledge they need when they encounter an obstacle. In this way, the learning-by-doing process would be made more effective. Users don’t get answers, but help is there if they need it.
This technology seems like a handy tool for a new employee, but what does the documentation process look like for a current employee? In this master-apprentice model, is the bot still watching?
The KNIC system would not constantly monitor the employee – it would be a choice whether or not the technology is used in a particular situation.
“Right now, graduate students or workers transitioning to a new job have to sit down and write down everything they know, and it’s a very difficult process. Instead, they could use our system to demonstrate the different tasks they perform and ask the system to ask about parts that are unclear or represent something new or unfamiliar,” Pujara explained. .
That way, it would be a “supportive tool in the same way someone might use Zoom to record a lecture or a demo,” Pujara added.
just the beginning
KNIC is still in its early stages, in fact, it’s only month six of what’s supposed to be a three-year project. Currently, the team is working on building a working prototype for the data science field.
One of the future challenges of this project is to be able to make it more generalizable.
“It’s one thing to follow someone doing a task on the computer when everything they do is on the computer. Whereas if a trainee cook accidentally puts egg yolks in place of egg whites, that’s a very different type of error and it’s very different to detect,” Boschee said.
On top of that, machines are still nowhere near as powerful as humans. We still have a long way to go before AI can completely replace the role of a colleague training someone new.
Jonathan May, Associate Professor of Computer Science Researchstated that although Al is “very weak compared to what humans can do”, ISI is able to work on “simpler problems which would further increase efficiency and allow for better knowledge transfer”.
One of the areas where this project is applicable, Boschee said, is in the armed forces.
“The government cares about this because it’s a very common scenario in a military context, because people are on rotation, and making those role transitions without needing to go to the previous type would be extremely helpful,” explained Boschee.
Pujara and Boschee run the broadcast system and interfaces. Muhao Chen, assistant research professor of computer science, develops the basic knowledge and the extraction element for knowledge capture.
May, along with David Traum, director of natural language research at the USC Viterbi Institute for Creative Technologies (ICT), are both working on designing interfaces and building AI agents that would interact with people. human user.
The KNIC project shows a promising future for effective knowledge transfer in the workplace. With an AI companion on your shoulder, tackling a new role can be less scary and smoother.
(1) KMASS seeks to create technology that facilitates the documentation, acquisition and application of knowledge. DARPA chose proposals from various research institutions, one of which is ISI’s KNIC project, which focuses on the field of data science. This fall, the project received a grant of almost seven million dollars.
Posted on January 31, 2023
Last updated on January 31, 2023