Cerebras: Streamlining Everyday Activities for Memory-Impaired Individuals

Assisting Individuals with Dementia Through Task Management

Millions of people around the world are affected by Alzheimer’s and dementia, conditions that significantly impair their ability to manage everyday tasks, remember important details, and maintain their independence. Tasks such as taking medication on time or checking blood sugar levels are critical for these individuals. However, the tools available to assist with these tasks, like static to-do lists or verbal reminders, often fall short. Static lists require constant updates, which can be challenging for someone experiencing cognitive difficulties, while verbal reminders are often too generic and easily forgotten.

The Impact of Inefficiencies

A major challenge in managing tasks for individuals with dementia has been the slow processing speed of task extraction technologies. Before advancements like Cerebras’ inference technology, it was nearly impossible to extract actionable tasks from conversations in real-time. Conventional systems often took over 300 milliseconds to process spoken commands into actionable tasks, creating two significant issues:

1. Break in Conversation Flow: For those dealing with memory loss, even a brief delay can disrupt the natural flow of conversation. This interruption often leads to frustration with the system as users either lose patience or forget the context of their task during the delay.
2. Incomplete or Missed Tasks: Slow processing speeds made it difficult for systems to keep up with fast-paced, dynamic conversations. This often resulted in critical tasks being skipped or only partially recorded, making these tools unreliable for real-world use.

   Consider a conversation like, "Hi Jensen, remember to pick up groceries and water the plants this afternoon." Previous solutions might have only partially extracted these tasks or processed them so slowly that users had to repeat or manually re-enter them, defeating the purpose of the tool.

   Without fast inference, systems like Memo-ry were ineffective as real-time companions. Slow task extraction reduced user trust and made it difficult to rely on these tools for daily life.

   The Solution
   

   Memo-ry addresses these challenges with Cerebras’ advanced inference technology, which processes speech into actionable tasks in nearly real-time. By reducing latency to under 50 milliseconds, Memo-ry provides a seamless user experience that allows for natural conversation flow without interruptions.

   For example, from a conversation like, "Hi Jensen, good morning. I’m so excited to see you again. Let’s go over today’s plan. First, you have a doctor’s appointment at 10:30 AM. After that, make sure to pick up your prescription from the pharmacy and don’t forget to water the plants this afternoon," Memo-ry instantly extracts:

   • Attend a doctor’s appointment at 10:30 AM
   • Pick up a prescription from the pharmacy
   • Water the plants

     The fast inference capabilities were essential for Memo-ry’s success, as real-time task extraction is crucial for individuals managing cognitive difficulties. Without it, task extraction would feel slow or incomplete, defeating the app’s purpose of simplifying daily management for people with dementia.

     Building the Solution
     

     Architecture Overview

     Memo-ry integrates several components to provide seamless real-time task extraction:

   • Frontend Interface: Built with HTML and CSS to ensure an accessible and user-friendly design.
   • Speech Recognition and Transcription: Utilizes the Web Speech API for real-time speech-to-text processing.
   • Task Extraction Engine: Leverages the Cerebras inference API to process transcripts into actionable tasks with sub-50ms latency, ensuring smooth task extraction without disrupting conversation flow.
   • Dynamic Task Display: Uses JavaScript to dynamically update the UI, presenting tasks as interactive checklists.

     Performance Comparison

     By employing Cerebras technology, task extraction latency dropped from around 300ms with traditional GPUs to under 50ms. This 85% reduction in latency significantly improved real-time interactions, which was essential for the app’s use case. Without Cerebras, it would have been challenging for the application to update tasks in real-time alongside conversations, substantially reducing its usability.

     Future Developments
     

     Memo-ry holds immense potential to enhance the lives of individuals with memory challenges, and future updates aim to expand its functionality:

   • Voice Feedback: Provide auditory reinforcement by reading tasks aloud to users.
   • Smart Home Integration: Enable compatibility with devices like Alexa and Google Home for hands-free use.
   • Mobile App Development: Transition to mobile platforms with push notifications and advanced accessibility features.
   • Caregiver Collaboration: Add features for remote task management, allowing caregivers to support users from anywhere.
   • AI Personalization: Tailor task suggestions based on user behavior, habits, and needs.

     These enhancements will make Memo-ry an indispensable tool for both individuals with memory loss and their caregivers.

     About the Creator
     

     The development of Memo-ry was made possible through the Cerebras Fellows Program, which emphasizes fast inference technology. This allowed the full potential of the app to be realized, bridging the gap between innovation and usability.

     Jensen, an engineering student at MIT, is passionate about using technology to empower underserved communities. With a background in problem-solving from a young age, Jensen has worked on significant projects such as a patent-pending algae biofiltration system to combat air pollution and an ultrasonic backpack for the visually impaired. At MIT, Jensen has expanded into fields like geospatial analysis, natural language processing, and social impact entrepreneurship, leading projects such as NASA-funded research on urban heat islands and studies on air quality’s effect on muon flux. As a low-income student, Jensen is also dedicated to increasing access to STEM education, teaching free coding classes in Boston, and creating robotics curricula for underserved schools.

     About the Fellows Program
     

     Cerebras inference technology is driving the next generation of AI applications, offering speeds 70 times faster than traditional GPUs. The Cerebras x Bain Capital Ventures Fellows Program invites engineers, researchers, and students to create impactful, next-level products powered by instant AI. For more information, visit Cerebras Fellows Program.

     Through this program, innovations like Memo-ry are not just possible—they are being actively developed to make a real-world impact. By continuing to explore and expand the capabilities of such applications, the future looks promising for individuals managing memory loss and related challenges.

For more Information, Refer to this article.