 |
|
|
|||||
Understanding intelligence and creating intelligent machines are grand scientific challenges of our times. The ability to learn from experience is a cornerstone of intelligence for machines and living beings alike. In a remarkably prescient 1948 report, Alan Turingthe father of modern computer scienceproposed the construction of machines that display intelligent behavior. He also discussed the education of such machines by means of rewards and punishments. Turings ideas ultimately led to the development of reinforcement learning, a branch of artificial intelligence. Reinforcement learning designs intelligent agents by training them to maximize rewards as they interact with their environment. As a machine learning researcher, I find it fitting that reinforcement learning pioneers Andrew Barto and Richard Sutton were awarded the 2024 ACM Turing Award. What is reinforcement learning? Animal trainers know that animal behavior can be influenced by rewarding desirable behaviors. A dog trainer gives the dog a treat when it does a trick correctly. This reinforces the behavior, and the dog is more likely to do the trick correctly the next time. Reinforcement learning borrowed this insight from animal psychology. But reinforcement learning is about training computational agents, not animals. The agent can be a software agent like a chess-playing program. But the agent can also be an embodied entity like a robot learning to do household chores. Similarly, the environment of an agent can be virtual, like the chessboard or the designed world in a video game. But it can also be a house where a robot is working. Just like animals, an agent can perceive aspects of its environment and take actions. A chess-playing agent can access the chessboard configuration and make moves. A robot can sense its surroundings with cameras and microphones. It can use its motors to move about in the physical world. Agents also have goals that their human designers program into them. A chess-playing agents goal is to win the game. A robots goal might be to assist its human owner with household chores. The reinforcement learning problem in AI is how to design agents that achieve their goals by perceiving and acting in their environments. Reinforcement learning makes a bold claim: All goals can be achieved by designing a numerical signal, called the reward, and having the agent maximize the total sum of rewards it receives. Researchers do not know if this claim is actually true, because of the wide variety of possible goals. Therefore, it is often referred to as the reward hypothesis. Sometimes it is easy to pick a reward signal corresponding to a goal. For a chess-playing agent, the reward can be +1 for a win, 0 for a draw, and -1 for a loss. It is less clear how to design a reward signal for a helpful household robotic assistant. Nevertheless, the list of applications where reinforcement learning researchers have been able to design good reward signals is growing. A big success of reinforcement learning was in the board game Go. Researchers thought that Go was much harder than chess for machines to master. The company DeepMind, now Google DeepMind, used reinforcement learning to create AlphaGo. AlphaGo defeated top Go player Lee Sedol in a five-match game in 2016. A more recent example is the use of reinforcement learning to make chatbots such as ChatGPT more helpful. Reinforcement learning is also being used to improve the reasoning capabilities of chatbots. Reinforcement learnings origins However, none of these successes could have been foreseen in the 1980s. That is when Barto and his then-PhD student Sutton proposed reinforcement learning as a general problem-solving framework. They drew inspiration not only from animal psychology but also from the field of control theory, the use of feedback to influence a systems behavior, and optimization, a branch of mathematics that studies how to select the best choice among a range of available options. They provided the research community with mathematical foundations that have stood the test of time. They also created algorithms that have now become standard tools in the field. It is a rare advantage for a field when pioneers take the time to write a textbook. Shining examples like The Nature of the Chemical Bond by Linus Pauling and The Art of Computer Programming by Donald E. Knuth are memorable because they are few and far between. Sutton and Bartos Reinforcement Learning: An Introduction was first published in 1998. A second edition came out in 2018. Their book has influenced a generation of researchers and has been cited more than 75,000 times. Reinforcement learning has also had an unexpected impact on neuroscience. The neurotransmitter dopamine plays a key role in reward-driven behaviors in humans and animals. Researchers have used specific algorithms developed in reinforcement learning to explain experimental findings in people and animals dopamine system. Barto and Suttons foundational work, vision and advocacy have helped reinforcement learning grow. Their work has inspired a large body of research, made an impact on real-world applications, and attracted huge investments by tech companies. Reinforcement learning researchers, Im sure, will continue to see further ahead by standing on their shoulders. Ambuj Tewari is a professor of statistics at the University of Michigan. This article is republished from The Conversation under a Creative Commons license. Read the original article.
Category:
E-Commerce
For Paul, a finance administrator, things came to a head when his report mistakenly included 7,000,000 of costs rather than 700,000. Fearing accusations of fraud, Paul disclosed his recent dementia diagnosis to his boss. Six weeks of sick leave became six months, and then a stepping stone to early retirement. Several years later, Paul regrets his unwanted unemployment, but at the time there didnt seem to be an alternative. Paul was participating in an unrelated study about public transport when he told us about his unemployment. As researchers, we had heard many similar accounts so we decided to dig down into the research on work and dementia. We were curious about how typical Pauls experience was of the trajectories of people diagnosed while working. The ageing of populations around the world is influencing our lives in many ways. More people are extending their working lives beyond traditional retirement ages, and many more are being diagnosed with dementia. Around 9% of the worlds 55 million people with dementia are under 65, with around 370,000 new cases of young-onset dementia annually. It is striking then, that despite government and business commitments to support longer working lives and inclusive employment practices, workers with dementia are largely ignored. What little evidence we have paints a picture of widespread and unwanted unemployment. For some, this takes the form of redundancy or retirement. For others like Paul a period of temporary leave gradually evolves into a permanent exit. Alongside workforce ageing, digital transformation is perhaps the single most important development in modern industry. Almost all our working lives are now shaped by digital technologies in some form. Older people are often stereotyped as technologically incompetent. This can be even worse for people with dementia. When exciting digital innovations are discussed in relation to them, the focus is almost always on providing care. But someone diagnosed with dementia in their 60s today might have been blogging in their 30s, scrolling social media on a phone in their 40s and using a smart home assistant in their 50s. The tech is here already The reality is that many people with dementia use digital tools every day. This ranges from familiar products like Google Maps to more cutting-edge technologies. A person with dementia recently introduced us to their voice-activated AI companion, with which they watch and discuss films. These companions can provide vital social interaction for people fearing judgement or isolation because of their cognitive decline. Far from being a barrier, digital technologies could offer ways to help people with dementia to enjoy positive working lives, just as they help workers who dont have dementia. The trick is to use them to tailor work and workplaces to the individual. For example, if a worker is struggling to remember appointments, automated and shared calendar scheduling can take care of that. If a worker has impaired wayfinding, mapping apps can be tailored to working environments and live location data can be used to guide staff around complicated sites. This is hardly futuristic tech. Many of us would struggle without our online calendars and maps. Research shows that touchscreens can be particularly challenging for older people with dementia. To make interfaces more suitable, developers could encourage the integration of voice-operated smart assistants into employee workstations (think of Amazons Alexa or Apples Siri). While discussions of dementia often focus on memory loss, the various types of dementia are associated with a wide range of symptoms. One very common symptom is the struggle to find the right words. But recent developments in generative AI (like OpenAIs ChatGPT) are proficient at predicting and expressing the next word in a sequence. These tools are also excellent at transforming text into different formats. Guidance on dementia-friendly information recommends features such as large fonts, single-clause sentences and single-syllable words. A generative AI tool could quickly transform documents into dementia-friendly formats. The integration of these tools into emailing and writing applications could make a lot of work far more accessible to people with dementia. These days, it makes little sense for workers to be manually entering costings into a spreadsheet. Dementia or no dementia, these practices are ripe for human error. By outsourcing them to digital technologies, we can free up our ageing workforces to use their unmatched skills, such as networking and experience. In practice, employers will likely be responsible for supporting positive working lives with dementia in the future. The best way to do this will be to develop strategies, in consultation with people with dementia, that identify interventions suitable for the workplace. Then, when an employee is diagnosed, they can pick and mix a personalised collection of tools to address their needs. Right now, we are not aware of any workplace that has such a strategy. But many organisations already have robust policies for other conditions. Our own employer, the University of Bath, has a repository of reasonable adjustments that can be tailored to support staff and students experiencing mental illness. Dementia could be approached in much the same way. The UK government is currently attempting to increase the number of people with disabilities participating in the labour market. It is simultaneously driving an agenda to increase the use of AI throughout the country. The potential of a digital working life for people with dementia highlights both promise and peril. Simply forcing every person into work is a surefire way of turnin challenging situations into real problems. But providing tailored support for those who want to work can enrich organisations and workers alike. James Fletcher is a lecturer (assistant professor) of management information, decisions & operations at the Institute for Digital Security and Behaviour at the University of Bath. Olivia Brown is an associate professor in digital futures at the University of Bath. This article is republished from The Conversation under a Creative Commons license. Read the original article.
Category:
E-Commerce
It would seem unlikely for clothing designers to get their wheels turning by thinking about what happens to garments when people are through with them, but thats exactly the sort of backward thinking that led to Under Armours new regenerative sportswear collection, created in collaboration with Portland, Oregon-based Unless Collective. The collection, which is making its debut in Italy this week during Milan Design Week, comprises footwear and clothing made entirely from plants and plant-based materials. That means theyre biodegradable and compostable. All of our products make good dirt, says Eric Liedtke, Unless cofounder and Under Armour EVP of brand strategy, who spoke with Fast Company from Milan. Unless was acquired by Baltimore-based Under Armour last summer, and Liedtke says that its allowed Unless to tap into Under Armours large base of resources and partnerships to expand its offerings and development operationsthe new regenerative collection is the result. Were here to introduce the idea of regenerative fashion,” he says. “What we mean is that things come from plants and minerals, natural materials, and then go back to being natural materials. . . . When youre regenerative, you add value back to the ecosystem, rather than being destructive. Liedtke says that 70% of clothing is created from petroleum-based feedstock, mostly various types of plastics, which never completely vanish or go awaythey break down into microplastics and end up in the food and water supply. But his new clothing line does break down and go away; once youre through with one of Unlesss garments, for instance, you can bury it in your backyard garden, and itll compost away. In an industrial composter, an Unless tee shirt will decompose within weeks. [Photos: Under Armour] The new collection features shoes, jackets, vests, shirts, and more that are made from a variety of plant materials. For instance, shoe liners and soles are made from coconut husks and natural rubber latex, buttons are made from corozo nuts, Kapok cotton is used for insulation in vests and jackets, while cotton remains a staple for shirts and other garments. Liedtke says that the garments are built to last, too, and could be compared to products from companies like Russell, Champion, Carhartt, or Dickies. And for those worried about their clothes decomposing while they sit in a dresser, he says not to worry: It takes very specific conditions to initiate the composting processconditions hopefully not present in the typical closet or bedroom. The collection is meant to be provocative, in some ways, and bring attention to the pollution that modern fashion and clothing manufacturing produces. In that way, its not too different from how companies like Beyond Meat and Impossible Meat disrupted the meat industry, or how EVs have shaken up the auto market in recent years. Liedtke hopes that at least some clothing manufacturers will follow suit and start using more natural materials, rather than plastics, to cut down on waste and pollution. The future is regenerative, he says. The question now is scaling it, and telling people about it.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||