The phrase “practice makes perfect” is usually reserved for humans, but it’s also a great maxim for robots newly deployed in unfamiliar environments.

Picture a robot arriving in a warehouse. It comes packaged with the skills it was trained on, like placing an object, and now it needs to pick items from a shelf it’s not familiar with. At first, the machine struggles with this, since it needs to get acquainted with its new surroundings. To improve, the robot will need to understand which skills within an overall task it needs improvement on, then specialize (or parameterize) that action.

A human onsite could program the robot to optimize its performance, but researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and The AI Institute have developed a more effective alternative. Presented at the Robotics: Science and Systems Conference last month, their “Estimate, Extrapolate, and Situate” (EES) algorithm enables these machines to practice on their own, potentially helping them improve at useful tasks in factories, households, and hospitals. 

Sizing up the situation

To help robots get better at activities like sweeping floors, EES works with a vision system that locates and tracks the machine’s surroundings. Then, the algorithm estimates how reliably the robot executes an action (like sweeping) and whether it would be worthwhile to practice more. EES forecasts how well the robot could perform the overall task if it refines that particular skill, and finally, it practices. The vision system subsequently checks whether that skill was done correctly after each attempt.

EES could come in handy in places like a hospital, factory, house, or coffee shop. For example, if you wanted a robot to clean up your living room, it would need help practicing skills like sweeping. According to Nishanth Kumar SM ’24 and his colleagues, though, EES could help that robot improve without human intervention, using only a few practice trials.

“Going into this project, we wondered if this specialization would be possible in a reasonable amount of samples on a real robot,” says Kumar, co-lead author of a paper describing the work, PhD student in electrical engineering and computer science, and a CSAIL affiliate. “Now, we have an algorithm that enables robots to get meaningfully better at specific skills in a reasonable amount of time with tens or hundreds of data points, an upgrade from the thousands or millions of samples that a standard reinforcement learning algorithm requires.”

See Spot sweep

EES’s knack for efficient learning was evident when implemented on Boston Dynamics’ Spot quadruped during research trials at The AI Institute. The robot, which has an arm attached to its back, completed manipulation tasks after practicing for a few hours. In one demonstration, the robot learned how to securely place a ball and ring on a slanted table in roughly three hours. In another, the algorithm guided the machine to improve at sweeping toys into a bin within about two hours. Both results appear to be an upgrade from previous frameworks, which would have likely taken more than 10 hours per task.

“We aimed to have the robot collect its own experience so it can better choose which strategies will work well in its deployment,” says co-lead author Tom Silver SM ’20, PhD ’24, an electrical engineering and computer science (EECS) alumnus and CSAIL affiliate who is now an assistant professor at Princeton University. “By focusing on what the robot knows, we sought to answer a key question: In the library of skills that the robot has, which is the one that would be most useful to practice right now?”

EES could eventually help streamline autonomous practice for robots in new deployment environments, but for now, it comes with a few limitations. For starters, they used tables that were low to the ground, which made it easier for the robot to see its objects. Kumar and Silver also 3D printed an attachable handle that made the brush easier for Spot to grab. The robot didn’t detect some items and identified objects in the wrong places, so the researchers counted those errors as failures.

Giving robots homework

The researchers note that the practice speeds from the physical experiments could be accelerated further with the help of a simulator. Instead of physically working at each skill autonomously, the robot could eventually combine real and virtual practice. They hope to make their system faster with less latency, engineering EES to overcome the imaging delays the researchers experienced. In the future, they may investigate an algorithm that reasons over sequences of practice attempts instead of planning which skills to refine.

“Enabling robots to learn on their own is both incredibly useful and extremely challenging,” says Danfei Xu, an assistant professor in the School of Interactive Computing at Georgia Tech and a research scientist at NVIDIA AI, who was not involved with this work. “In the future, home robots will be sold to all sorts of households and expected to perform a wide range of tasks. We can’t possibly program everything they need to know beforehand, so it’s essential that they can learn on the job. However, letting robots loose to explore and learn without guidance can be very slow and might lead to unintended consequences. The research by Silver and his colleagues introduces an algorithm that allows robots to practice their skills autonomously in a structured way. This is a big step towards creating home robots that can continuously evolve and improve on their own.”

Silver and Kumar’s co-authors are The AI Institute researchers Stephen Proulx and Jennifer Barry, plus four CSAIL members: Northeastern University PhD student and visiting researcher Linfeng Zhao, MIT EECS PhD student Willie McClinton, and MIT EECS professors Leslie Pack Kaelbling and Tomás Lozano-Pérez. Their work was supported, in part, by The AI Institute, the U.S. National Science Foundation, the U.S. Air Force Office of Scientific Research, the U.S. Office of Naval Research, the U.S. Army Research Office, and MIT Quest for Intelligence, with high-performance computing resources from the MIT SuperCloud and Lincoln Laboratory Supercomputing Center.

You’ve likely met someone who identifies as a visual or auditory learner, but others absorb knowledge through a different modality: touch. Being able to understand tactile interactions is especially important for tasks such as learning delicate surgeries and playing musical instruments, but unlike video and audio, touch is difficult to record and transfer.

To tap into this challenge, researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and elsewhere developed an embroidered smart glove that can capture, reproduce, and relay touch-based instructions. To complement the wearable device, the team also developed a simple machine-learning agent that adapts to how different users react to tactile feedback, optimizing their experience. The new system could potentially help teach people physical skills, improve responsive robot teleoperation, and assist with training in virtual reality.

An open-access paper describing the work was published in Nature Communications on Jan. 29.

Will I be able to play the piano?

To create their smart glove, the researchers used a digital embroidery machine to seamlessly embed tactile sensors and haptic actuators (a device that provides touch-based feedback) into textiles. This technology is present in smartphones, where haptic responses are triggered by tapping on the touch screen. For example, if you press down on an iPhone app, you’ll feel a slight vibration coming from that specific part of your screen. In the same way, the new adaptive wearable sends feedback to different parts of your hand to indicate optimal motions to execute different skills.

The smart glove could teach users how to play the piano, for instance. In a demonstration, an expert was tasked with recording a simple tune over a section of keys, using the smart glove to capture the sequence by which they pressed their fingers to the keyboard. Then, a machine-learning agent converted that sequence into haptic feedback, which was then fed into the students’ gloves to follow as instructions. With their hands hovering over that same section, actuators vibrated on the fingers corresponding to the keys below. The pipeline optimizes these directions for each user, accounting for the subjective nature of touch interactions.

“Humans engage in a wide variety of tasks by constantly interacting with the world around them,” says Yiyue Luo MS ’20, lead author of the paper, PhD student in MIT’s Department of Electrical Engineering and Computer Science (EECS), and CSAIL affiliate. “We don’t usually share these physical interactions with others. Instead, we often learn by observing their movements, like with piano-playing and dance routines.

“The main challenge in relaying tactile interactions is that everyone perceives haptic feedback differently,” adds Luo. “This roadblock inspired us to develop a machine-learning agent that learns to generate adaptive haptics for individuals’ gloves, introducing them to a more hands-on approach to learning optimal motion.”

The wearable system is customized to fit the specifications of a user’s hand via a digital fabrication method. A computer produces a cutout based on individuals’ hand measurements, then an embroidery machine stitches the sensors and haptics in. Within 10 minutes, the soft, fabric-based wearable is ready to wear. Initially trained on 12 users’ haptic responses, its adaptive machine-learning model only needs 15 seconds of new user data to personalize feedback.

In two other experiments, tactile directions with time-sensitive feedback were transferred to users sporting the gloves while playing laptop games. In a rhythm game, the players learned to follow a narrow, winding path to bump into a goal area, and in a racing game, drivers collected coins and maintained the balance of their vehicle on their way to the finish line. Luo’s team found that participants earned the highest game scores through optimized haptics, as opposed to without haptics and with unoptimized haptics.

“This work is the first step to building personalized AI agents that continuously capture data about the user and the environment,” says senior author Wojciech Matusik, MIT professor of electrical engineering and computer science and head of the Computational Design and Fabrication Group within CSAIL. “These agents then assist them in performing complex tasks, learning new skills, and promoting better behaviors.”

Bringing a lifelike experience to electronic settings

In robotic teleoperation, the researchers found that their gloves could transfer force sensations to robotic arms, helping them complete more delicate grasping tasks. “It’s kind of like trying to teach a robot to behave like a human,” says Luo. In one instance, the MIT team used human teleoperators to teach a robot how to secure different types of bread without deforming them. By teaching optimal grasping, humans could precisely control the robotic systems in environments like manufacturing, where these machines could collaborate more safely and effectively with their operators.

“The technology powering the embroidered smart glove is an important innovation for robots,” says Daniela Rus, the Andrew (1956) and Erna Viterbi Professor of Electrical Engineering and Computer Science at MIT, CSAIL director, and author on the paper. “With its ability to capture tactile interactions at high resolution, akin to human skin, this sensor enables robots to perceive the world through touch. The seamless integration of tactile sensors into textiles bridges the divide between physical actions and digital feedback, offering vast potential in responsive robot teleoperation and immersive virtual reality training.”

Likewise, the interface could create more immersive experiences in virtual reality. Wearing smart gloves would add tactile sensations to digital environments in video games, where gamers could feel around their surroundings to avoid obstacles. Additionally, the interface would provide a more personalized and touch-based experience in virtual training courses used by surgeons, firefighters, and pilots, where precision is paramount.

While these wearables could provide a more hands-on experience for users, Luo and her group believe they could extend their wearable technology beyond fingers. With stronger haptic feedback, the interfaces could guide feet, hips, and other body parts less sensitive than hands.

Luo also noted that with a more complex artificial intelligence agent, her team’s technology could assist with more involved tasks, like manipulating clay or driving an airplane. Currently, the interface can only assist with simple motions like pressing a key or gripping an object. In the future, the MIT system could incorporate more user data and fabricate more conformal and tight wearables to better account for how hand movements impact haptic perceptions.

Luo, Matusik, and Rus authored the paper with EECS Microsystems Technology Laboratories Director and Professor Tomás Palacios; CSAIL members Chao Liu, Young Joong Lee, Joseph DelPreto, Michael Foshey, and professor and principal investigator Antonio Torralba; Kiu Wu of LightSpeed Studios; and Yunzhu Li of the University of Illinois at Urbana-Champaign.

The work was supported, in part, by an MIT Schwarzman College of Computing Fellowship via Google and a GIST-MIT Research Collaboration grant, with additional help from Wistron, Toyota Research Institute, and Ericsson.

In order for natural language to be an effective form of communication, the parties involved need to be able to understand words and their context, assume that the content is largely shared in good faith and is trustworthy, reason about the information being shared, and then apply it to real-world scenarios. MIT PhD students interning with the MIT-IBM Watson AI Lab — Athul Paul Jacob SM ’22, Maohao Shen SM ’23, Victor Butoi, and Andi Peng SM ’23 — are working to attack each step of this process that’s baked into natural language models, so that the AI systems can be more dependable and accurate for users.

To achieve this, Jacob’s research strikes at the heart of existing natural language models to improve the output, using game theory. His interests, he says, are two-fold: “One is understanding how humans behave, using the lens of multi-agent systems and language understanding, and the second thing is, ‘How do you use that as an insight to build better AI systems?’” His work stems from the board game “Diplomacy,” where his research team developed a system that could learn and predict human behaviors and negotiate strategically to achieve a desired, optimal outcome.

“This was a game where you need to build trust; you need to communicate using language. You need to also play against six other players at the same time, which were very different from all the kinds of task domains people were tackling in the past,” says Jacob, referring to other games like poker and GO that researchers put to neural networks. “In doing so, there were a lot of research challenges. One was, ‘How do you model humans? How do you know whether when humans tend to act irrationally?’” Jacob and his research mentors — including Associate Professor Jacob Andreas and Assistant Professor Gabriele Farina of the MIT Department of Electrical Engineering and Computer Science (EECS), and the MIT-IBM Watson AI Lab’s Yikang Shen — recast the problem of language generation as a two-player game.

Using “generator” and “discriminator” models, Jacob’s team developed a natural language system to produce answers to questions and then observe the answers and determine if they are correct. If they are, the AI system receives a point; if not, no point is rewarded. Language models notoriously tend to hallucinate, making them less trustworthy; this no-regret learning algorithm collaboratively takes a natural language model and encourages the system’s answers to be more truthful and reliable, while keeping the solutions close to the pre-trained language model’s priors. Jacob says that using this technique in conjunction with a smaller language model could, likely, make it competitive with the same performance of a model many times bigger.  

Once a language model generates a result, researchers ideally want its confidence in its generation to align with its accuracy, but this frequently isn’t the case. Hallucinations can occur with the model reporting high confidence when it should be low. Maohao Shen and his group, with mentors Gregory Wornell, Sumitomo Professor of Engineering in EECS, and lab researchers with IBM Research Subhro Das, Prasanna Sattigeri, and Soumya Ghosh — are looking to fix this through uncertainty quantification (UQ). “Our project aims to calibrate language models when they are poorly calibrated,” says Shen. Specifically, they’re looking at the classification problem. For this, Shen allows a language model to generate free text, which is then converted into a multiple-choice classification task. For instance, they might ask the model to solve a math problem and then ask it if the answer it generated is correct as “yes, no, or maybe.” This helps to determine if the model is over- or under-confident.

Automating this, the team developed a technique that helps tune the confidence output by a pre-trained language model. The researchers trained an auxiliary model using the ground-truth information in order for their system to be able to correct the language model. “If your model is over-confident in its prediction, we are able to detect it and make it less confident, and vice versa,” explains Shen. The team evaluated their technique on multiple popular benchmark datasets to show how well it generalizes to unseen tasks to realign the accuracy and confidence of language model predictions. “After training, you can just plug in and apply this technique to new tasks without any other supervision,” says Shen. “The only thing you need is the data for that new task.”

Victor Butoi also enhances model capability, but instead, his lab team — which includes John Guttag, the Dugald C. Jackson Professor of Computer Science and Electrical Engineering in EECS; lab researchers Leonid Karlinsky and Rogerio Feris of IBM Research; and lab affiliates Hilde Kühne of the University of Bonn and Wei Lin of Graz University of Technology — is creating techniques to allow vision-language models to reason about what they’re seeing, and is designing prompts to unlock new learning abilities and understand key phrases.

Compositional reasoning is just another aspect of the decision-making process that we ask machine-learning models to perform in order for them to be helpful in real-world situations, explains Butoi. “You need to be able to think about problems compositionally and solve subtasks,” says Butoi, “like, if you’re saying the chair is to the left of the person, you need to recognize both the chair and the person. You need to understand directions.” And then once the model understands “left,” the research team wants the model to be able to answer other questions involving “left.”

Surprisingly, vision-language models do not reason well about composition, Butoi explains, but they can be helped to, using a model that can “lead the witness”, if you will. The team developed a model that was tweaked using a technique called low-rank adaptation of large language models (LoRA) and trained on an annotated dataset called Visual Genome, which has objects in an image and arrows denoting relationships, like directions. In this case, the trained LoRA model would be guided to say something about “left” relationships, and this caption output would then be used to provide context and prompt the vision-language model, making it a “significantly easier task,” says Butoi.

In the world of robotics, AI systems also engage with their surroundings using computer vision and language. The settings may range from warehouses to the home. Andi Peng and mentors MIT’s H.N. Slater Professor in Aeronautics and Astronautics Julie Shah and Chuang Gan, of the lab and the University of Massachusetts at Amherst, are focusing on assisting people with physical constraints, using virtual worlds. For this, Peng’s group is developing two embodied AI models — a “human” that needs support and a helper agent — in a simulated environment called ThreeDWorld. Focusing on human/robot interactions, the team leverages semantic priors captured by large language models to aid the helper AI to infer what abilities the “human” agent might not be able to do and the motivation behind actions of the “human,” using natural language. The team’s looking to strengthen the helper’s sequential decision-making, bidirectional communication, ability to understand the physical scene, and how best to contribute.

“A lot of people think that AI programs should be autonomous, but I think that an important part of the process is that we build robots and systems for humans, and we want to convey human knowledge,” says Peng. “We don’t want a system to do something in a weird way; we want them to do it in a human way that we can understand.”