A new computer algorithm developed by Parham Aarabi of the University of Toronto can strategically store and recall information, just like our brains do. Edward S. Rogers Sr., associate professor in the Department of Electrical and Computer Engineering in the College of Applied S

A new computer algorithm developed by Parham Aarabi of the University of Toronto can store and recall information strategically, just like our brains do.

Edward S. Rogers Sr., associate professor in the Department of Electrical and Computer Engineering in the College of Applied Science and Engineering, has also developed an experimental tool that uses a new algorithm to help people with amnesia.

In the past, computers relied on users to tell them what information they should store. However, with the rise of artificial intelligence (AI) technology, deep learning and neural network , there has been a trend towards "fuzzy" methods.

Ten years ago, computing was absolute, CPU processed and stored in-memory data with precision to make binary decisions. There is no ambiguity.

now wants the computer to make an approximate conclusion and guess the percentage. The image processor is expected to tell us that, for example, there is a 10% chance that a photo contains a car and a 40% chance that it contains a pedestrian.

Aarabi extends the same fuzzy approach to storing and retrieving information by replicating several attributes that help humans determine what to remember and, just as importantly, what to forget.

Research shows that we tend to prioritize what happened more recently than what happened recently. It also emphasizes memories that are more important to us, condensing long narratives down to their essence.

For example, remembering to send my daughter to school today, promising to pay someone back, promising that I would read a research paper, but not remembering every second of my experience.

's ability to ignore certain information may significantly enhance existing machine learning models.

Today, machine learning algorithms scour millions of database entries, looking for patterns that will help them correctly associate a given input with a given output. Only after countless iterations does the algorithm eventually become accurate enough to handle new problems it hasn't seen yet.

If biologically inspired artificial memory enabled these algorithms to highlight the most relevant data they might arrive at meaningful results faster.

This method can also support natural language processing to help people with amnesia record key information.

Aarabi and his team have built such a tool using a simple email-based interface. It reminds participants of important information based on algorithmic priority and relevant indexing of keywords.

Ultimately, it targets memory loss. This helps them remember things in a very human, very gentle way without overwhelming them. Most task management tools are too complex to be useful in these situations.

The goal is to put the demo into people's hands, whether they're dealing with severe memory loss or just everyday stress, and see what feedback they get. The next steps are to build partnerships in health care to test in a more comprehensive way.

These days, artificial intelligence applications are increasingly appearing in many human-centered fields, and Professor Alabi hopes to ensure the potential of artificial intelligence in society by studying how to better integrate artificial intelligence with these soft fields. be fully utilized.

Aarabi says this algorithm is just the beginning, and that biologically inspired memory will likely bring artificial intelligence closer to human-level abilities.