Later On

A blog written for those whose interests more or less match mine.

Archive for the ‘Games’ Category

AlphaZero and the Protected Passed Pawn

leave a comment »

I thought this commentary was particularly good, though I realize tastes differ. Fascinating game.

Written by LeisureGuy

12 January 2018 at 11:16 pm

How to solve a crossword puzzle

leave a comment »

Some interesting strategies I didn’t know about, plus some useful knowledge—e.g., the Monday NY Times crossword puzzle is the easiest of the week, then they get progressively harder each day through Saturday, the hardest puzzle of the week. Sunday puzzle is about a Wednesday/Thursday level of difficulty.

Written by LeisureGuy

18 December 2017 at 10:16 am

Posted in Daily life, Games

Nail-biting snooker frame. Ronnie O’Sullivan is amazing.

leave a comment »

Listen to the audio, if you can. It really adds a lot as you hear how excited the announcers are getting.

Written by LeisureGuy

8 December 2017 at 4:26 pm

Posted in Games, Video

The NHL (and the NFL) have much to answer for: ‘I Have No Idea How to Tell This Horror Story’

leave a comment »

The NHL and NFL care only about the money they make and care nothing about the health of their players. From the NY Times:

John Branch, a sports reporter at The New York Times, has been in sporadic contact via text messaging and email with Walter Peat since writing about him and his son Stephen, a former N.H.L. player.

At the time of the article, in June 2016, Stephen Peat was 36 and experiencing debilitating headaches and violent mood swings. Peat was primarily an enforcer, a player designated to drop his gloves and square off in fist-to-fist combat with an opponent. The Peats presumed that Stephen’s problems were rooted in brain trauma sustained on the ice in so many fights.

Walter Peat, the head saw filer at a lumber mill in a suburb of Vancouver, British Columbia, gave The Times permission to publish the texts and emails he had sent John over the previous 18 months. Some have been trimmed for concision.

(Stephen disputes his father’s accounts and said, “I am disappointed in my father since I once held him so high on a pedestal.”)

JUNE 7, 2016

Again, we thank you for doing our story, which hasn’t ended yet. I just hope Stephen can get some help, going forward. We had a great day yesterday, as just the 2 of us went boating for the day. Beautiful day, making memories together.

DEC. 30, 2016: John is copied on an email to a doctor with the N.H.L. Players’ Association.

Dr. Rizos,

We are at a critical point in Stephen’s health. I am afraid, Stephen may become another statistic in NHL players who’s life is ended due to brain injuries suffered from playing. We are desperate for help, as we have run out of resources and energy to cope with this. The NHL has offered zero help. I am sorry we haven’t contacted you sooner, but Stephen is stubborn, and proud. It has gone past that, and hope someone will help. He is suffering badly from memory loss, depression, extreme headaches, and at times suicide thoughts. Along with this, tough when he get frustrated and anger comes out. He has got violent a couple times, and at times I am afraid for my life. I love my son very much, and you have no idea how much this hurts to see him like this. Please help.

JUNE 1, 2017

Just an update on Stephen, not sure if the NHL might want to know, but Stephen is in lockup, been arrested 2 times in the last week for parole violation. May spend the next 6 months in jail, he is real bad shape, and like I said before, I don’t have the resources or the knowledge to deal with this. I saw him yesterday, looks horrible, he is homeless. I know one can lead a horse to water theory, but I am afraid this could be very close to his end. At times he has no idea who he is or where he is.

Stephen needs special medical attention badly, as I also believe he has gone back to self medicating, not sure, but you can’t imagine what bad shape he is in. I fear he may not make it out of lock up, but there is nothing I can do, or don’t know what to do. The system is flawed.

He is in lock up now, most likely won’t get released now. Only violation is not seeing his parole officer, but he forgets, a complete mess right now. Probably the safest place for him, but he won’t get fixed in a cement cell.

JUNE 7, 2017

I will be honest, my health has suffered thru this, and I am at a financial crisis as Stephen has gone through $120,000 since getting out of rehab. And he is demanding more every day, to a point if I continue this, I will be on the street. Stephen can’t comprehend my financial stress he has put on me. I went to meet him at Clover Towing the other day, Stephen looked in real bad shape, as he didn’t even know I was there. I had to leave, and stop in a parking lot next door and just broke down. The police were called by the towing company and they took him to jail. He should be in a hospital, not a jail, we can talk more about this. I will say I have had to get a no contact order as I fear for my safety now. If he is in a state where he doesn’t know who I am, it scares me.

JUNE 14, 2017 . . .

Continue reading.

Written by LeisureGuy

21 November 2017 at 3:02 pm

The Uncanny Resurrection of Dungeons & Dragons

leave a comment »

Neima Jahromi writes in the New Yorker:

The clinical psychologist Jon Freeman was feeling burnt out. He spent his days at a corporate office in Manhattan, managing dozens of research assistants as they tested pharmaceuticals on people with anxiety, depression, and insomnia. Looking for an escape hatch, he noticed that his daughter often had nothing to do after school. She would pick up her Nintendo Wii controller and drift “into this world of digital isolation,” Freeman recalled. From time to time, he enticed her back into social existence with board games. “Then I had this idea: Couldn’t we do this on a larger scale? Could we expand this to our neighborhood?”
Freeman quit his job, and, shortly thereafter, in 2011, the first customers—initially, his daughter’s friends—arrived at his pop-up board-game club and café, Brooklyn Strategist, a place where children and their parents could sit down and play games, both classic and obscure, over veggie platters and homemade ginger ale. Looking back at his work in the research lab, he paired cognitive-ability tests with the board games that he had on hand, and divided these amusements by brain function—kids worked their way around their frontal lobes a die roll at a time.
One day, a child who had grown tired of a sports-statistics game asked if Freeman had heard of the role-playing game Dungeons & Dragons, and if they could play it. The game has no board and no cards. Occasionally, players make use of maps. At its best, it’s a story told between the players, who control characters (elves, dwarves, gnomes, humans), and the Dungeon Master, who describes the world and uses dice to determine outcomes in the second person (“You come across a band of orcs, travelling down the road. What do you do?”). Freeman refused for a week or two—the game was too open-ended, and didn’t have a straightforward cognitive benefit—but the customer persisted, so he went up into his parents’ attic, dug out all his old D. & D. manuals, and wrote an adventure. “I tried to give them a little flavor of everything,” he told me, “A little dungeon crawl, a little fighting monsters. They ate it up.” Word got out. A few months later, a parent stopped him on the street with tears in her eyes. “What are you guys doing?” she asked him. Her son was dyslexic and had been role-playing at Brooklyn Strategist for a couple of weeks. Before D. & D., he couldn’t focus on writing for more than a few seconds. Now he was staying up all night to draft stories about his character. “Whatever it is, bottle it and sell it to me,” the mother said.
Freeman got a permanent space in 2012 and added French-press coffee. A few months later, Gygax, a once defunct magazine named for the Dungeons & Dragons co-creator Gary Gygax, chose Brooklyn Strategist to host its relaunch party. A reporter for Wired, covering the event, asked the magazine’s founders why they wanted to waste their energy on such a publication (not to mention such a store) when “it’s video games, not Dungeons and Dragons and other RPGs, that are getting all the attention?” This attention, it seems, has shifted. Two popular role-playing shows, “The Adventure Zone” and “Critical Role,” sent Freeman’s older patrons to their knees, begging for more D. & D. time in the store. Soon, Freeman had to hire half a dozen paid Dungeon Masters for the kids and has now begun training volunteer Dungeon Masters to guide adventures for the adults who drop in on Thursdays to fight goblins, trick castle guards, and drink wine.
Dungeons & Dragons nights have spread into classrooms and game stores across the country. Forty dollars in Portland, Oregon, gets you into Orcs! Orcs! Orcs!, a “Tavern-inspired” pop-up restaurant with D. & D. games and artisanal delicacies. (One night, it boasted “tankards of beer” and “a whole roast pig.”) In Massachusetts, snow or shine, a series of role-playing camps called Guard Up offers children the chance to chase each other through the fields of Burlington with foam swords and Nerf blasters, while somehow also learning. (Each summer, in one camp, novels like “Animal Farm” or “Twenty Thousand Leagues Under the Sea” are adapted into a mock zombie apocalypse that is then played out by the campers. In another, at a moment of detente, Gandalf might appear on the edge of a running track to give physics lessons.) “I’ve had parents get very upset with me,” said Freeman, who recently opened another store near Columbia University. “Because they sign their kids up for role playing and my staff is trying to expand their horizons beyond D. & D. and into other independent games. But the parents are, like, ‘If they can’t play D. & D., then I don’t know if this is going to work.’ ”
This turn of events might shock a time traveller from the twentieth century. In the seventies and eighties, Dungeons & Dragons, with its supernatural themes, became the fixation of an overheated news media in the midst of a culture war. Role players were seen as closet cases, the least productive kind of geek, retreating to basements to open maps, spill out bags of dice, and light candles by which to see their medieval figurines. They squared with no one. Unlike their hippie peers, they had dropped out without bothering to tune in. On the other side of politics,Christian moralists’ cries of the occult and anxiety about witchcraft followed D. & D. players everywhere. Worse still, parents feared how this enveloping set of lies about druids in dark cloaks and paladins on horseback could tip already vulnerable minds off the cliff of reality. At the end of the 1982 TV movie “Mazes and Monsters,” a troubled gamer, played by a pre-fame Tom Hanks, loses touch and starts to believe that he really does live beside an evil wood in need of heroes. “He saw the monsters. We did not,” his ex-girlfriend says in a voice-over. “We saw nothing but the death of hope, and the loss of our friend.”
Decades passed, D. & D. movies and cartoons came and went, and the game remade itself over and over. But interest fell like an orc beneath a bastard sword. The game’s designers, surrounded by copycats and perplexed about how to bring D. & D. online, made flat-footed attempts at developing new rule books to mimic the video games that D. & D. had inspired. Gygax died, in 2008, occasioning a wealth of tributes but little enthusiasm. Then, a fifth edition of D. & D. rules came out, in 2014, and, somehow, the culture was receptive again to bags of holding and silver-haired drow. People started buying up these volumes in droves. “More people are interested in D&D than we thought,” the game’s lead developer, Mike Mearls, said, as print runs repeatedly sold out. “Who are these people? What do they want?”
In 2017, gathering your friends in a room, setting your devices aside, and taking turns to contrive a story that exists largely in your head gives off a radical whiff for a completely different reason than it did in 1987. And the fear that a role-playing game might wound the psychologically fragile seems to have flipped on its head. Therapists use D. & D. to get troubled kids to talk about experiences that might otherwise embarrass them, and children with autism use the game to improve their social skills. Last year, researchers found that a group of a hundred and twenty-seven role players exhibited above-average levels of empathy, and a Brazilian study from 2013 showed that role-playing classes were an extremely effective way to teach cellular biology to medical undergraduates.
Adult D. & D. acolytes are everywhere now, too. The likes of Drew Barrymore and Vin Diesel regularly take up the twenty-sided die (or at least profess to do so). Tech workers from Silicon Valley to Brooklyn have long-running campaigns, and the showrunners and the novelist behind “Game of Thrones” have all been Dungeon Masters. (It’s also big with comedy improvisers in Los Angeles, but it’s no surprise that theatre kids have nerdy hobbies.) Nevertheless, the image of the recluse persists even among fans. “We’re going to alienate ninety-nine per cent of the people out there right now,” Stephen Colbert told Anderson Cooper last year, on “The Late Show,” as they fondly recalled their respective turns as an elven thief and a witch. “The shut-in at home is really excited,” Cooper replied. “Neckbeards,” Colbert added.
The “neckbeards” may be more numerous now than he and Cooper realize. “The Big Bang Theory” is a sitcom about young scientists at CalTech who spend most of their time shuttling between their laboratories and the comic-book store. The show’s protagonists also play a lot of D. & D. In one episode, a theoretical physicist takes on the guise of the Dungeon Master to relieve a microbiologist of her distress over the restraints of her pregnancy. She pretends, for an evening, to live in a world where only men are with child (“Your husband is home trying not to pee when he laughs”), to drink ale out of the skull of a goblin, and to eat sushi made from the meat of a monster that she has butchered herself. Fourteen million people tuned in.
Dungeons & Dragons seems to have been waiting for us somewhere under the particular psyche of this generation, a psyche that may have been coaxed into fantasy mania by the media that surrounded it. . .

Continue reading.

Written by LeisureGuy

24 October 2017 at 12:56 pm

AlphaGo Zero: Learning from scratch

leave a comment »

A very good (and very interesting) post at DeepMind.com:

Artificial intelligence research has made rapid progress in a wide variety of domains from speech recognition and image classification to genomics and drug discovery. In many cases, these are specialist systems that leverage enormous amounts of human expertise and data.

However, for some problems this human knowledge may be too expensive, too unreliable or simply unavailable. As a result, a long-standing ambition of AI research is to bypass this step, creating algorithms that achieve superhuman performance in the most challenging domains with no human input. In our most recent paper, published in the journal Nature, we demonstrate a significant step towards this goal.

The paper introduces AlphaGo Zero, the latest evolution of AlphaGo, the first computer program to defeat a world champion at the ancient Chinese game of Go. Zero is even more powerful and is arguably the strongest Go player in history.

Previous versions of AlphaGo initially trained on thousands of human amateur and professional games to learn how to play Go. AlphaGo Zero skips this step and learns to play simply by playing games against itself, starting from completely random play. In doing so, it quickly surpassed human level of play and defeated the previously published champion-defeating version of AlphaGo by 100 games to 0.

It is able to do this by using a novel form of reinforcement learning, in which AlphaGo Zero becomes its own teacher. The system starts off with a neural network that knows nothing about the game of Go. It then plays games against itself, by combining this neural network with a powerful search algorithm. As it plays, the neural network is tuned and updated to predict moves, as well as the eventual winner of the games.

This updated neural network is then recombined with the search algorithm to create a new, stronger version of AlphaGo Zero, and the process begins again. In each iteration, the performance of the system improves by a small amount, and the quality of the self-play games increases, leading to more and more accurate neural networks and ever stronger versions of AlphaGo Zero.

This technique is more powerful than previous versions of AlphaGo because it is no longer constrained by the limits of human knowledge. Instead, it is able to learn tabula rasa from the strongest player in the world: AlphaGo itself.

It also differs from previous versions in other notable ways.

  • AlphaGo Zero only uses the black and white stones from the Go board as its input, whereas previous versions of AlphaGo included a small number of hand-engineered features.
  • It uses one neural network rather than two. Earlier versions of AlphaGo used a “policy network” to select the next move to play and a ”value network” to predict the winner of the game from each position. These are combined in AlphaGo Zero, allowing it to be trained and evaluated more efficiently.
  • AlphaGo Zero does not use “rollouts” – fast, random games used by other Go programs to predict which player will win from the current board position. Instead, it relies on its high quality neural networks to evaluate positions.

All of these differences help improve the performance of the system and make it more general. But it is the algorithmic change that makes the system much more powerful and efficient.

After just three days of self-play training, AlphaGo Zero emphatically defeated the previously published version of AlphaGo – which had itself defeated 18-time world champion Lee Sedol – by 100 games to 0. After 40 days of self training, AlphaGo Zero became even stronger, outperforming the version of AlphaGo known as “Master”, which has defeated the world’s best players and world number one Ke Jie. . .

Continue reading.

Written by LeisureGuy

22 October 2017 at 2:14 pm

Posted in Go, Software, Technology

One more step toward the Singularity: Artificial Intelligence Learns to Learn Entirely on Its Own

leave a comment »

In science-fiction about the Singularity, the point at which AI could instruct itself was a big step, leading to a rapid increase in capabilities as a result of positive feedback. Reality is not, of course, science-fiction, but the report of AI instructing itself does catch the eye.

Kevin Hartnett writes in Quanta:

A mere 19 months after dethroning the world’s top human Go player, the computer program AlphaGo has smashed an even more momentous barrier: It can now achieve unprecedented levels of mastery purely by teaching itself. Starting with zero knowledge of Go strategy and no training by humans, the new iteration of the program, called AlphaGo Zero, needed just three days to invent advanced strategies undiscovered by human players in the multi-millennia history of the game. By freeing artificial intelligence from a dependence on human knowledge, the breakthrough removes a primary limit on how smart machines can become.

Earlier versions of AlphaGo were taught to play the game using two methods. In the first, called supervised learning, researchers fed the program 100,000 top amateur Go games and taught it to imitate what it saw. In the second, called reinforcement learning, they had the program play itself and learn from the results.

AlphaGo Zero skipped the first step. The program began as a blank slate, knowing only the rules of Go, and played games against itself. At first, it placed stones randomly on the board. Over time it got better at evaluating board positions and identifying advantageous moves. It also learned many of the canonical elements of Go strategy and discovered new strategies all its own. “When you learn to imitate humans the best you can do is learn to imitate humans,” said Satinder Singh, a computer scientist at the University of Michigan who was not involved with the research. “In many complex situations there are new insights you’ll never discover.”

After three days of training and 4.9 million training games, the researchers matched AlphaGo Zero against the earlier champion-beating version of the program. AlphaGo Zero won 100 games to zero.

To expert observers, the rout was stunning. Pure reinforcement learning would seem to be no match for the overwhelming number of possibilities in Go, which is vastly more complex than chess: You’d have expected AlphaGo Zero to spend forever searching blindly for a decent strategy. Instead, it rapidly found its way to superhuman abilities.

The efficiency of the learning process owes to a feedback loop. Like its predecessor, AlphaGo Zero determines what move to play through a process called a “tree search.” The program starts with the current board and considers the possible moves. It then considers what moves its opponent could play in each of the resulting boards, and then the moves it could play in response and so on, creating a branching tree diagram that simulates different combinations of play resulting in different board setups.

AlphaGo Zero can’t follow every branch of the tree all the way through, since that would require inordinate computing power. Instead, it selectively prunes branches by deciding which paths seem most promising. It makes that calculation — of which paths to prune — based on what it has learned in earlier play about the moves and overall board setups that lead to wins.

Earlier versions of AlphaGo did all this, too. What’s novel about AlphaGo Zero is that instead of just running the tree search and making a move, it remembers the outcome of the tree search — and eventually of the game. It then uses that information to update its estimates of promising moves and the probability of winning from different positions. As a result, the next time it runs the tree search it can use its improved estimates, trained with the results of previous tree searches, to generate even better estimates of the best possible move.

The computational strategy that underlies AlphaGo Zero is effective primarily in situations in which you have an extremely large number of possibilities and want to find the optimal one. In the Nature paper describing the research, the authors of AlphaGo Zero suggest that their system could be useful in materials exploration — where you want to identify atomic combinations that yield materials with different properties — and protein folding, where you want to understand how a protein’s precise three-dimensional structure determines its function.

As for Go,  . . .

Continue reading.

Written by LeisureGuy

18 October 2017 at 4:12 pm

Posted in Go, Software, Technology

%d bloggers like this: