Experts at MIT have created a unique memory system. While humans may forget, robots will simply retrieve the forgotten item.

Experts at MIT have created a unique memory system. While humans may forget, robots will simply retrieve the forgotten item.

      MIT’s new robotic memory could make your lost keys the robot's responsibility

      Robots might become the ideal companion for forgetful individuals. Researchers at MIT have created a long-term memory framework for robots that enables them to construct a detailed mental representation of vast, complex environments. The system, known as DAAAM (Describe Anything, Anywhere, Anytime, at Any Moment), aims to help robots retain information about objects, locations, and specifics over time.

      While this may not seem particularly exciting, robots often struggle with something that humans handle effortlessly. For instance, you might recall seeing your keys on the kitchen counter last night or remember that a half-finished component was left in a factory bin. In contrast, a robot working alongside you would find it challenging to associate that object with its location in a meaningful way.

      A map that robots can truly understand

      DAAAM seeks to address this issue by merging two existing technologies utilized by robots: computer vision and 3D mapping. As a robot navigates through its surroundings, it assigns detailed language descriptions to the objects it observes and saves them within a spatial map. Thus, instead of merely recognizing that an object exists at a certain coordinate, the robot could remember that there’s a red bicycle with a flat tire near a specific building or that a particular tool was located in a designated work area.

      As a result, a user could pose questions like, “Where did I leave my wallet?” or “Retrieve the component we began assembling last night,” and the robot could search its memory for the relevant object and its location.

      Not yet ready for your home…

      MIT’s DAAAM operates swiftly enough for a mobile robot to use in real-time scenarios. The researchers also discovered that it provides more accurate responses to queries compared to current methods, depending on the type of question. However, this capability won’t be available in your robot vacuum anytime soon. This research was shared at the Conference on Computer Vision and Pattern Recognition, and the study is accessible as a preprint.

      Researchers are still focused on enhancements, such as improving the system's confidence levels and assisting it in recalling significant events within an environment. For the time being, the concept is intriguing. At present, AI is a trending topic, but practical intelligence that effectively serves real-world needs certainly sounds more appealing.

      Vikhyaat Vivek is a tech journalist and reviewer with seven years of experience covering consumer hardware, focusing on…

      This new video editor allows Claude to organize, generate, and edit directly on your timeline

      For years, AI video tools have largely functioned outside the editing workflow. You create a clip, download it, import it into your editing software, and continue from there. A newly introduced app, Palmier Pro, aims to streamline this process by integrating AI directly into the video timeline.

      This software, available for macOS, is marketed as a video editor that can be operated by Claude. Instead of treating AI as a standalone chatbot or content generator, Palmier is designed for an AI assistant to interact with an active video project and make modifications within it.

      A peculiar little electric nose may be the key for smart fridges

      Researchers at UC Berkeley have developed an electric nose capable of detecting gases associated with spoiled food and common allergens more reliably than a human sniff test. This device utilizes a 16-sensor gas sensor chip that converts reactions to food-related gases into electrical signals.

      Assessing food safety can be challenging, as items don’t always appear or smell dangerous before they become problematic. Milk, eggs, chicken, fruit, and nuts emit different chemical signatures, and individuals typically have to rely on what their noses detect in the moment.

      Samsung’s pet tech needs just a picture to identify health issues affecting your furry friends

      This is the first mainstream smartphone-based system that monitors pet health without the need for specialized hardware.

      Samsung has integrated AI into various areas, from assessing your sleep quality to enhancing your TV viewing experience and monitoring your refrigerator contents. At VivaTech 2026 in Paris, the company unveiled a personal and beneficial pet health feature that employs AI to identify potential health concerns before they lead to costly veterinary bills.

Experts at MIT have created a unique memory system. While humans may forget, robots will simply retrieve the forgotten item. Experts at MIT have created a unique memory system. While humans may forget, robots will simply retrieve the forgotten item. Experts at MIT have created a unique memory system. While humans may forget, robots will simply retrieve the forgotten item. Experts at MIT have created a unique memory system. While humans may forget, robots will simply retrieve the forgotten item. Experts at MIT have created a unique memory system. While humans may forget, robots will simply retrieve the forgotten item. Experts at MIT have created a unique memory system. While humans may forget, robots will simply retrieve the forgotten item. Experts at MIT have created a unique memory system. While humans may forget, robots will simply retrieve the forgotten item.

Other articles

Within SURBL, the email blacklist that verifies your links rather than your IP address. Within SURBL, the email blacklist that verifies your links rather than your IP address. SURBL identifies the URLs within your emails instead of the sender's IP address. A single link to a flagged domain can quietly deactivate every link in a sent message, and most senders remain unaware of this occurrence. Apple allows competing app stores on iOS in Brazil. Apple allows competing app stores on iOS in Brazil. Apple will permit alternative app marketplaces and third-party payments in Brazil with the release of iOS 26.5, as a result of a CADE antitrust agreement. These modifications reflect concessions similar to those made in the EU. Architect Labs secures $24 million for the design of custom AI chips. Architect Labs secures $24 million for the design of custom AI chips. Architect Labs secured $24 million in seed funding led by Kindred Ventures to develop an AI that designs and validates custom chips, positioning itself against Broadcom and Marvell. AI vision is becoming increasingly resource-intensive, and this approach helps to reduce its appetite. AI vision is becoming increasingly resource-intensive, and this approach helps to reduce its appetite. KAIST's Upsample Anything addresses the memory challenge associated with enhanced on-device AI vision by recovering high-resolution visual details from compressed image data, enabling smartphones to avoid processing everything at full resolution initially. UPS is creating a real-time digital replica of its complete logistics network, refreshed every 10 minutes. UPS is creating a real-time digital replica of its complete logistics network, refreshed every 10 minutes. UPS outlined its AI initiatives, which include a digital twin of its global network, autonomous control towers for customer use, and RFID-enabled tracking for all vehicles in the US. Waymo's robotaxis continue to encounter new obstacles, with construction zones being the most recent. Waymo's robotaxis continue to encounter new obstacles, with construction zones being the most recent. Waymo's robotaxis navigated highway construction zones 13 times before the company decided to withdraw the entire fleet from highway operations.

Experts at MIT have created a unique memory system. While humans may forget, robots will simply retrieve the forgotten item.

Researchers at MIT have created an AI memory system enabling robots to retain information about objects, locations, and details from real-world settings over time.