Feed aggregator
Citing Trump tariffs, Canadian province eliminates a carbon tax
Three NASA satellites are dying. Their end could disrupt climate data.
California’s snowpack data likely signals another fire-prone summer
Climate firm that partnered with Meta, Microsoft goes bankrupt
Japan’s $1.7T pension fund offers new backing to ESG
Jesuit prefers prison over fine to draw attention to climate change
Researchers teach LLMs to solve complex planning challenges
Imagine a coffee company trying to optimize its supply chain. The company sources beans from three suppliers, roasts them at two facilities into either dark or light coffee, and then ships the roasted coffee to three retail locations. The suppliers have different fixed capacity, and roasting costs and shipping costs vary from place to place.
The company seeks to minimize costs while meeting a 23 percent increase in demand.
Wouldn’t it be easier for the company to just ask ChatGPT to come up with an optimal plan? In fact, for all their incredible capabilities, large language models (LLMs) often perform poorly when tasked with directly solving such complicated planning problems on their own.
Rather than trying to change the model to make an LLM a better planner, MIT researchers took a different approach. They introduced a framework that guides an LLM to break down the problem like a human would, and then automatically solve it using a powerful software tool.
A user only needs to describe the problem in natural language — no task-specific examples are needed to train or prompt the LLM. The model encodes a user’s text prompt into a format that can be unraveled by an optimization solver designed to efficiently crack extremely tough planning challenges.
During the formulation process, the LLM checks its work at multiple intermediate steps to make sure the plan is described correctly to the solver. If it spots an error, rather than giving up, the LLM tries to fix the broken part of the formulation.
When the researchers tested their framework on nine complex challenges, such as minimizing the distance warehouse robots must travel to complete tasks, it achieved an 85 percent success rate, whereas the best baseline only achieved a 39 percent success rate.
The versatile framework could be applied to a range of multistep planning tasks, such as scheduling airline crews or managing machine time in a factory.
“Our research introduces a framework that essentially acts as a smart assistant for planning problems. It can figure out the best plan that meets all the needs you have, even if the rules are complicated or unusual,” says Yilun Hao, a graduate student in the MIT Laboratory for Information and Decision Systems (LIDS) and lead author of a paper on this research.
She is joined on the paper by Yang Zhang, a research scientist at the MIT-IBM Watson AI Lab; and senior author Chuchu Fan, an associate professor of aeronautics and astronautics and LIDS principal investigator. The research will be presented at the International Conference on Learning Representations.
Optimization 101
The Fan group develops algorithms that automatically solve what are known as combinatorial optimization problems. These vast problems have many interrelated decision variables, each with multiple options that rapidly add up to billions of potential choices.
Humans solve such problems by narrowing them down to a few options and then determining which one leads to the best overall plan. The researchers’ algorithmic solvers apply the same principles to optimization problems that are far too complex for a human to crack.
But the solvers they develop tend to have steep learning curves and are typically only used by experts.
“We thought that LLMs could allow nonexperts to use these solving algorithms. In our lab, we take a domain expert’s problem and formalize it into a problem our solver can solve. Could we teach an LLM to do the same thing?” Fan says.
Using the framework the researchers developed, called LLM-Based Formalized Programming (LLMFP), a person provides a natural language description of the problem, background information on the task, and a query that describes their goal.
Then LLMFP prompts an LLM to reason about the problem and determine the decision variables and key constraints that will shape the optimal solution.
LLMFP asks the LLM to detail the requirements of each variable before encoding the information into a mathematical formulation of an optimization problem. It writes code that encodes the problem and calls the attached optimization solver, which arrives at an ideal solution.
“It is similar to how we teach undergrads about optimization problems at MIT. We don’t teach them just one domain. We teach them the methodology,” Fan adds.
As long as the inputs to the solver are correct, it will give the right answer. Any mistakes in the solution come from errors in the formulation process.
To ensure it has found a working plan, LLMFP analyzes the solution and modifies any incorrect steps in the problem formulation. Once the plan passes this self-assessment, the solution is described to the user in natural language.
Perfecting the plan
This self-assessment module also allows the LLM to add any implicit constraints it missed the first time around, Hao says.
For instance, if the framework is optimizing a supply chain to minimize costs for a coffeeshop, a human knows the coffeeshop can’t ship a negative amount of roasted beans, but an LLM might not realize that.
The self-assessment step would flag that error and prompt the model to fix it.
“Plus, an LLM can adapt to the preferences of the user. If the model realizes a particular user does not like to change the time or budget of their travel plans, it can suggest changing things that fit the user’s needs,” Fan says.
In a series of tests, their framework achieved an average success rate between 83 and 87 percent across nine diverse planning problems using several LLMs. While some baseline models were better at certain problems, LLMFP achieved an overall success rate about twice as high as the baseline techniques.
Unlike these other approaches, LLMFP does not require domain-specific examples for training. It can find the optimal solution to a planning problem right out of the box.
In addition, the user can adapt LLMFP for different optimization solvers by adjusting the prompts fed to the LLM.
“With LLMs, we have an opportunity to create an interface that allows people to use tools from other domains to solve problems in ways they might not have been thinking about before,” Fan says.
In the future, the researchers want to enable LLMFP to take images as input to supplement the descriptions of a planning problem. This would help the framework solve tasks that are particularly hard to fully describe with natural language.
This work was funded, in part, by the Office of Naval Research and the MIT-IBM Watson AI Lab.
Looking under the hood at the brain’s language system
As a young girl growing up in the former Soviet Union, Evelina Fedorenko PhD ’07 studied several languages, including English, as her mother hoped that it would give her the chance to eventually move abroad for better opportunities.
Her language studies not only helped her establish a new life in the United States as an adult, but also led to a lifelong interest in linguistics and how the brain processes language. Now an associate professor of brain and cognitive sciences at MIT, Fedorenko studies the brain’s language-processing regions: how they arise, whether they are shared with other mental functions, and how each region contributes to language comprehension and production.
Fedorenko’s early work helped to identify the precise locations of the brain’s language-processing regions, and she has been building on that work to generate insight into how different neuronal populations in those regions implement linguistic computations.
“It took a while to develop the approach and figure out how to quickly and reliably find these regions in individual brains, given this standard problem of the brain being a little different across people,” she says. “Then we just kept going, asking questions like: Does language overlap with other functions that are similar to it? How is the system organized internally? Do different parts of this network do different things? There are dozens and dozens of questions you can ask, and many directions that we have pushed on.”
Among some of the more recent directions, she is exploring how the brain’s language-processing regions develop early in life, through studies of very young children, people with unusual brain architecture, and computational models known as large language models.
From Russia to MIT
Fedorenko grew up in the Russian city of Volgograd, which was then part of the Soviet Union. When the Soviet Union broke up in 1991, her mother, a mechanical engineer, lost her job, and the family struggled to make ends meet.
“It was a really intense and painful time,” Fedorenko recalls. “But one thing that was always very stable for me is that I always had a lot of love, from my parents, my grandparents, and my aunt and uncle. That was really important and gave me the confidence that if I worked hard and had a goal, that I could achieve whatever I dreamed about.”
Fedorenko did work hard in school, studying English, French, German, Polish, and Spanish, and she also participated in math competitions. As a 15-year-old, she spent a year attending high school in Alabama, as part of a program that placed students from the former Soviet Union with American families. She had been thinking about applying to universities in Europe but changed her plans when she realized the American higher education system offered more academic flexibility.
After being admitted to Harvard University with a full scholarship, she returned to the United States in 1998 and earned her bachelor’s degree in psychology and linguistics, while also working multiple jobs to send money home to help her family.
While at Harvard, she also took classes at MIT and ended up deciding to apply to the Institute for graduate school. For her PhD research at MIT, she worked with Ted Gibson, a professor of brain and cognitive sciences, and later, Nancy Kanwisher, the Walter A. Rosenblith Professor of Cognitive Neuroscience. She began by using functional magnetic resonance imaging (fMRI) to study brain regions that appeared to respond preferentially to music, but she soon switched to studying brain responses to language.
She found that working with Kanwisher, who studies the functional organization of the human brain but hadn’t worked much on language before, helped Fedorenko to build a research program free of potential biases baked into some of the early work on language processing in the brain.
“We really kind of started from scratch,” Fedorenko says, “combining the knowledge of language processing I have gained by working with Gibson and the rigorous neuroscience approaches that Kanwisher had developed when studying the visual system.”
After finishing her PhD in 2007, Fedorenko stayed at MIT for a few years as a postdoc funded by the National Institutes of Health, continuing her research with Kanwisher. During that time, she and Kanwisher developed techniques to identify language-processing regions in different people, and discovered new evidence that certain parts of the brain respond selectively to language. Fedorenko then spent five years as a research faculty member at Massachusetts General Hospital, before receiving an offer to join the faculty at MIT in 2019.
How the brain processes language
Since starting her lab at MIT’s McGovern Institute for Brain Research, Fedorenko and her trainees have made several discoveries that have helped to refine neuroscientists’ understanding of the brain’s language-processing regions, which are spread across the left frontal and temporal lobes of the brain.
In a series of studies, her lab showed that these regions are highly selective for language and are not engaged by activities such as listening to music, reading computer code, or interpreting facial expressions, all of which have been argued to be share similarities with language processing.
“We’ve separated the language-processing machinery from various other systems, including the system for general fluid thinking, and the systems for social perception and reasoning, which support the processing of communicative signals, like facial expressions and gestures, and reasoning about others’ beliefs and desires,” Fedorenko says. “So that was a significant finding, that this system really is its own thing.”
More recently, Fedorenko has turned her attention to figuring out, in more detail, the functions of different parts of the language processing network. In one recent study, she identified distinct neuronal populations within these regions that appear to have different temporal windows for processing linguistic content, ranging from just one word up to six words.
She is also studying how language-processing circuits arise in the brain, with ongoing studies in which she and a postdoc in her lab are using fMRI to scan the brains of young children, observing how their language regions behave even before the children have fully learned to speak and understand language.
Large language models (similar to ChatGPT) can help with these types of developmental questions, as the researchers can better control the language inputs to the model and have continuous access to its abilities and representations at different stages of learning.
“You can train models in different ways, on different kinds of language, in different kind of regimens. For example, training on simpler language first and then more complex language, or on language combined with some visual inputs. Then you can look at the performance of these language models on different tasks, and also examine changes in their internal representations across the training trajectory, to test which model best captures the trajectory of human language learning,” Fedorenko says.
To gain another window into how the brain develops language ability, Fedorenko launched the Interesting Brains Project several years ago. Through this project, she is studying people who experienced some type of brain damage early in life, such as a prenatal stroke, or brain deformation as a result of a congenital cyst. In some of these individuals, their conditions destroyed or significantly deformed the brain’s typical language-processing areas, but all of these individuals are cognitively indistinguishable from individuals with typical brains: They still learned to speak and understand language normally, and in some cases, they didn’t even realize that their brains were in some way atypical until they were adults.
“That study is all about plasticity and redundancy in the brain, trying to figure out what brains can cope with, and how” Fedorenko says. “Are there many solutions to build a human mind, even when the neural infrastructure is so different-looking?”
Vote for “How to Fix the Internet” in the Webby Awards People's Voice Competition!
EFF’s “How to Fix the Internet” podcast is a nominee in the Webby Awards 29th Annual People's Voice competition – and we need your support to bring the trophy home!
We keep hearing all these dystopian stories about technology’s impact on our lives and our futures — from tracking-based surveillance capitalism to the dominance of a few large platforms choking innovation to the growing pressure by authoritarian governments to control what we see and say. The landscape can feel bleak. Exposing and articulating these problems is important, but so is envisioning and then building a better future.
That’s where our podcast comes in. Through curious conversations with some of the leading minds in law and technology, “How to Fix the Internet” explores creative solutions to some of today’s biggest tech challenges.
Over our five seasons, we’ve had well-known, mainstream names like Marc Maron to discuss patent trolls, Adam Savage to discuss the rights to tinker and repair, Dave Eggers to discuss when to set technology aside, and U.S. Sen. Ron Wyden, D-OR, to discuss how Congress can foster an internet that benefits everyone. But we’ve also had lesser-known names who do vital, thought-provoking work – Taiwan’s then-Minister of Digital Affairs Audrey Tang discussed seeing democracy as a kind of open-source social technology, Alice Marwick discussed the spread of conspiracy theories and disinformation, Catherine Bracy discussed getting tech companies to support (not exploit) the communities they call home, and Chancey Fleet discussing the need to include people with disabilities in every step of tech development and deployment.
We’ve just recorded our first interview for Season 6, and episodes should start dropping next month! Meanwhile, you can catch up on our past seasons to become deeply informed on vital technology issues and join the movement working to build a better technological future.
And if you’ve liked what you’ve heard, please throw us a vote in the Webbys competition!
Our deepest thanks to all our brilliant guests, and to the Alfred P. Sloan Foundation's Program in Public Understanding of Science and Technology, without whom this podcast would not be possible.
Click below to listen to the show now, or choose your podcast player:
%3Ciframe%20scrolling%3D%22no%22%20seamless%3D%22%22%20src%3D%22https%3A%2F%2Fplayer.simplecast.com%2F1c515ea8-cb6d-4f72-8d17-bc9b7a566869%3Fdark%3Dfalse%26amp%3Bshow%3Dtrue%22%20width%3D%22100%25%22%20height%3D%22480px%22%20frameborder%3D%22no%22%20allow%3D%22autoplay%22%3E%3C%2Fiframe%3E Privacy info. This embed will serve content from simplecast.comOr get our YouTube playlist! Or, listen to the episodes on the Internet Archive!
Trump names urologist as RFK’s deputy
Cell Phone OPSEC for Border Crossings
I have heard stories of more aggressive interrogation of electronic devices at US border crossings. I know a lot about securing computers, but very little about securing phones.
Are there easy ways to delete data—files, photos, etc.—on phones so it can’t be recovered? Does resetting a phone to factory defaults erase data, or is it still recoverable? That is, does the reset erase the old encryption key, or just sever the password that access that key? When the phone is rebooted, are deleted files still available?
We need answers for both iPhones and Android phones. And it’s not just the US; the world is going to become a more dangerous place to oppose state power...