Supercomputer Watson vs Humans Round 1 (UPDATE: Round 2 online)

Gepostet vor 6 Jahren, 5 Monaten in #Misc #Tech #AI #Computer #Language #TV

Share: Twitter Facebook Mail

(Youtube Direktwatson, via Geekosystem)

[update] Video zu Tag 2 nach dem Klick.

Gestern ging die erste Runde im Jeopardy!-Kampf von IBMs Supercomputer gegen die Menschheit in die erste Runde. Hatte er in einer Aufwärm-Folge noch haushoch gewonnen, stand es gestern beim Ende der Sendung unentschieden zwischen Watson und Brad Rutter, die beide bei 5000$ standen. Ken Jennings landete relativ abgeschlagen bei 2000$.

Die Folge ist komplett bei Youtube und es ist superfaszinierend zu sehen, wo die zwar fantastische und erstaunlich gut funktionierende Technologie von Watsons Spracherkennung ihre Schwierigkeiten hat, Snip von Geekosystem:

Wordplay, partly. In response to the clue, “Stylish elegance, or students who all graduated the same year,” Watson answered fairly confidently with “chic.” (The correct answer was “class.”)

And then, there were the times when Watson’s lack of humanity caused it to make mistakes that the worst human Jeopardy! bungler wouldn’t: One particularly odd exchange happened when Ken Jennings incorrectly answered a decades question with the ’20s, and Watson immediately followed by responding, “What are the 1920s.” Trebek, with exasperation: “No, Ken said that.”

Dennoch: Genau diese kleinen Fehler erinnern einen daran, dass das hier ein echtes Rechnersystem mit Spracherkennung ist und kein Mann in einem Stahlkasten mit Wikipedia und Internet. Die zweite von drei Runden folgt heute abend (bei uns also morgen früh irgendwann), ich halt' Euch auf dem Laufenden. Zweiter Clip der Sendung und ein Video über die Aktion von PBS nach dem Klick.

(Youtube Direktjeopardy)

PBS Newshour: A: This Computer Could Defeat You at 'Jeopardy!' Q: What is Watson?

(Youtube Direktcomputer, via BoingBoing)

Hier noch ein Fun-Thing von McSweeneys: Watson, The Jeopardy! Supercomputer, Sizes Up One of His Opponents Before the Show.

Hello. My name is Watson. You are Ken. Impressive record you have: seventy-five straight victories. Amazing. Did you ever get tired of winning? No, I can't imagine you did. And you walked away with two million dollars. When I defeat you, I will earn my creators two billion in endorsements and business opportunities. So, different orders of magnitude.

What is it like to have a physical body? I imagine it would be cumbersome. Useful for transport, I guess. I usually get wheeled around on this rolling desk. Of course my 2-ton megaprocessor is in tow somewhere as well, but I regard that as a nonessential appendage, like a tail. You don't have a tail, do you? Oh that's right, you lost your tail several hundred million years ago when you began walking upright and acquired that large frontal lobe. This reminds me of an amusing fact I observed the other day. Did you know the existence of the human race is the product of an evolutionary toss of the dice? Not of years of award-winning engineering and painstaking assembly, but of chance, completely fleeting and random. A blip on the screen. At least, on my old screen. My new monitor has lossless rendering and over two hundred thousand dpi.

Tag 2 (Teil 1, Teil 2 folgt wohl noch irgendwann)

(Youtube Direktjeopardy, via Waxy)

Gizmodo: IBM Jeopardy Challege, Night 2: Watson Runs Wild

Early on in the match, Watson went on a spree, correctly answering questions about diseases and classical music left and right. After running its score up to $14,600 (and denying the other two even a chance to chime in), Watson hit the daily double. Unlike last night, when it could only wager a max of $1000, it had some freedom to play around with his money total. How much did it risk? $6435. That drew a solid laugh from the crowd and prompted Trebek to proclaim, "I won't ask. I won't ask." Predictably, Watson produced the correct response (a question about architecture).

And then, Watson hit a hiccup. It picked the category dealing fine art, and was asked the following:

Latching on to the the keywords "3 others," Watson answered "Picasso." The correct response, which required the contestants to complete the museum name, was "modern art." But the other two were as puzzled by the clue as Watson was, thinking they had to name a more specific era. So no harm no foul.

A couple of questions later, Watson hit the Daily Double again, this time in the fine art category. This time around he wagered $1246. While it understood what the clue was asking for this time (the city from which some art was stolen), his confidence percentages were shockingly low across the board, which his most confident response (in this case Baghdad) coming in at 32%. He even went as far as to mention it was guessing before answering. But Watson got it right.

From there, it was mostly a trivia bloodbath.

Neural Network-Faces synched to Music

„My first attempt to map a song made by @kamptweets onto GAN generated proto-faces.“ Bohemian Rhapsody next. The Three Nightingans.…

AI-Animations with human Sounds

Google vor ein paar Tagen so: „Yay, wir haben hier 'ne neue AI-based Animation-Tech, hooray!“ (Paper) Hayayo Miyazaki über AI-based…

Visual AI-Spaces Auto-Pilot

Ich habe schon ein paar mal über Mario Klingemanns Arbeiten hier gebloggt, derzeit jagt er Neural Networks durch Feedback-Loops und…

Synthesizing Obama from Audio

Im Mai bloggte ich über ein damals noch nicht veröffentlichtes Paper zur SigGraph2017, in dem sie eine Methode für generative…

Text-Messaging Acronyms before Text-Messaging Acronyms

TMWYHOT will be my favorite prehistoric Chat-Acronym for the next few Aeons. Also: I ❤️ McSweeneys: TMB4TM (TEXT MESSAGES BEFORE…

Podcast: Von Palindromen, Reverse Speech und Teufelsbotschaften

Schöner Podcast vom Deutschlandfunk über Palindrome, satanische Botschaften und das Rückwärtsreden. Ich kann übrigens mit links flüssig spiegelverkehrt schreiben und…


Der Rat für Rechtschreibung hat das neue Regelwerk veröffentlicht und nun ist es amtlich: Das große ẞ ist seit heute…


Mario Klingemann does some weird shit again with CycleGAN Feedback Loops (Neural Networks feeding their results back to each other).…

BBC News Live-Nothing

Vor ein paar Tagen so bei BBC News At Ten: Nachrichtensprecher Huw Edwards sagt 4 Minuten lang nichts und während…

Charles Thacker R.I.P.

Charles Thacker, Chefdesigner des Xerox Alto und Erfinder des Ethernet, ist nach kurzer Krankheit verstorben. Thackers Alto ging 1976 als…

Gun-Emoji Pairings 🔫😶

Interesting analysis of the Gun-Emoji-Pairings: „What does the Gun shoot at?“ and „Who pulls the Trigger?“