| Both sides previous revision
Previous revision
Next revision
|
Previous revision
|
speed_of_ai_transition:range_of_human_performance:the_range_of_human_intelligence [2022/10/18 19:03]
harlanstewart |
speed_of_ai_transition:range_of_human_performance:the_range_of_human_intelligence [2023/09/11 21:38] (current)
jeffreyheninger The schemata weren't transferred over from the original. |
| |
| <HTML> | <HTML> |
| <p>The range of goldfish intelligence seems large relative to the space below it, as measured by performance on tasks we care about—despite the fact that human brains are extremely similar to each other.<span style="line-height: 1.5;"> </span></p> | <p>The range of human intelligence seems large relative to the space below it, as measured by performance on tasks we care about—despite the fact that human brains are extremely similar to each other.<span style="line-height: 1.5;"> </span></p> |
| </HTML> | </HTML> |
| |
| </HTML> | </HTML> |
| |
| | {{:speed_of_ai_transition:range_of_human_performance:sophvi_einstein-copy1.jpg?direct&400|}} |
| |
| |
| </HTML> | </HTML> |
| |
| | {{:speed_of_ai_transition:range_of_human_performance:modsofvi_einstein-copy.jpg?direct&400|}} |
| |
| | {{:speed_of_ai_transition:range_of_human_performance:vi_einstein-copy.jpg?direct&400|}} |
| |
| |
| </HTML> | </HTML> |
| |
| | 1. **Chess**// (See [[speed_of_ai_transition:range_of_human_performance:time_for_ai_to_cross_the_human_performance_range_in_chess|Time for AI to cross the human performance range in chess]])//: human chess Elo ratings conservatively range from around [[http://en.wikipedia.org/wiki/Elo_rating_system#United_States_Chess_Federation_ratings|800 (beginner)]] to 2800 (world champion). The following figure illustrates how it took chess AI roughly forty years to move incrementally from 1300 to 2800. |
| <HTML> | |
| <p>1. <strong>Chess:</strong> human chess Elo ratings conservatively range from around <a href="http://en.wikipedia.org/wiki/Elo_rating_system#United_States_Chess_Federation_ratings">800 (beginner)</a> to 2800 (world champion). The following figure illustrates how it took chess AI roughly forty years to move incrementally from 1300 to 2800.</p> | |
| </HTML> | |
| |
| |
| |
| |
| <HTML> | 2. **Go**// (See [[speed_of_ai_transition:range_of_human_performance:time_for_ai_to_cross_the_human_performance_range_in_go|Time for AI to cross the human performance range in Go]])//: Human go ratings [[http://en.wikipedia.org/wiki/Go_ranks_and_ratings|range]] from 30-20 kyu (beginner) to at least 9p (10p is a special title). Note that the numbers go downwards through kyu levels, then upward through dan levels, then upward through p(rofessional dan) levels. The following figure suggests that it took around 25 years for AI to cover most of this space (the top ratings seem to be [[http://en.wikipedia.org/wiki/Go_ranks_and_ratings#Elo-like_rating_systems_as_used_in_Go|closer together]] than the lower ones, though there are apparently [[https://en.wikipedia.org/wiki/Go_ranks_and_ratings#Winning_probabilities|multiple systems]] which vary). |
| <p>2. <strong>Go:</strong> Human go ratings <a href="http://en.wikipedia.org/wiki/Go_ranks_and_ratings">range</a> from 30-20 kyu (beginner) to at least 9p (10p is a special title). Note that the numbers go downwards through kyu levels, then upward through dan levels, then upward through p(rofessional dan) levels. The following figure suggests that it took around 25 years for AI to cover most of this space (the top ratings seem to be <a href="http://en.wikipedia.org/wiki/Go_ranks_and_ratings#Elo-like_rating_systems_as_used_in_Go">closer together</a> than the lower ones, though there are apparently <a href="https://en.wikipedia.org/wiki/Go_ranks_and_ratings#Winning_probabilities">multiple systems</a> which vary).</p> | |
| </HTML> | |
| |
| |
| </HTML> | </HTML> |
| |
| | 3. **StarCraft** //(See [[speed_of_ai_transition:range_of_human_performance:time_for_ai_to_cross_the_human_range_in_starcraft|Time for AI to cross the human range in StarCraft]]):// Beginner level StarCraft-playing AI seems to have been possible since the game was released in 1998. In 2018, AlphaStar beat a professional StarCraft player, suggesting that state-of-the-art StarCraft-playing AI had reached professional level. No AI seems to have above superhuman skills in StarCraft. |
| |
| <HTML> | 4. **Checkers** //(See [[speed_of_ai_transition:range_of_human_performance:time_for_ai_to_cross_the_human_range_in_english_draughts|Time for AI to cross the human range in English draughts]]):// According to [[http://en.wikipedia.org/wiki/Timeline_of_artificial_intelligence|Wikipedia’s timeline of AI]], a program was written in 1952 that could challenge a respectable amateur. In 1994 Chinook beat the second highest rated player ever. (In 2007 checkers was solved.) Thus it took around forty years to pass from amateur to world-class checkers-playing. We know nothing about whether intermediate progress was incremental however. |
| <p>3. <strong>Checkers: </strong>According to <a href="http://en.wikipedia.org/wiki/Timeline_of_artificial_intelligence">Wikipedia’s timeline of AI</a>, a program was written in 1952 that could challenge a respectable amateur. In 1994 Chinook beat the second highest rated player ever. (In 2007 checkers was solved.) Thus it took around forty years to pass from amateur to world-class checkers-playing. We know nothing about whether intermediate progress was incremental however.</p> | |
| </HTML> | 5. **Diagnosing diabetic retinopathy** //(See [[speed_of_ai_transition:range_of_human_performance:time_for_ai_to_cross_the_human_performance_range_in_diabetic_retinopathy|Time for AI to cross the human performance range in diabetic retinopathy]]))//: Diabetic retinopathy, a complication of diabetes in which high blood sugar levels damage the back of the eye, is diagnosed by examining images of the back of the eye. The first algorithm for detecting diabetic retinopathy, developed in 1996, could not detect it as well as ophthalmologists. In 2016, Google released results for algorithms that could detect diabetic retinopathy about as well as ophthalmologists. |
| |
| | 6. **Image classification** //(See [[featured_articles:time_for_ai_to_cross_the_human_performance_range_in_imagenet_image_classification|Time for AI to cross the human performance range in ImageNet image classification]])//: The first image classifiers were developed in 1998. ImageNet is a large collection of images organized into a hierarchy of noun categories, released in 2009. Our beginner level benchmark for ImageNet classification was first surpassed in 2012, by AlexNet. In 2015, an image classifier surpassed our high human benchmark, suggesting that state-of-the-art AI could classify images about as well as expert humans. |
| |
| <HTML> | <HTML> |
| <p>4. <strong>Physical manipulation:</strong> we have not investigated this much, but our impression is that robots are somewhere in the the fumbling and slow part of the human spectrum on <a href="http://youtu.be/oD9DE0HjMM4">some</a> <a href="https://www.youtube.com/watch?v=IBY4t8XxH7E">tasks</a>, and that nobody expects them to reach the ‘normal human abilities’ part any time soon (<a href="http://aiimpacts.wpengine.com/hanson-ai-expert-survey/" title="Hanson AI Expert Survey">Aaron Dollar estimates</a> robotic grasping manipulation in general is less than one percent of the way to human level from where it was 20 years ago).</p> | <p>7. <strong>Physical manipulation:</strong> we have not investigated this much, but our impression is that robots are somewhere in the the fumbling and slow part of the human spectrum on <a href="http://youtu.be/oD9DE0HjMM4">some</a> <a href="https://www.youtube.com/watch?v=IBY4t8XxH7E">tasks</a>, and that nobody expects them to reach the ‘normal human abilities’ part any time soon (<a href="http://aiimpacts.wpengine.com/hanson-ai-expert-survey/" title="Hanson AI Expert Survey">Aaron Dollar estimates</a> robotic grasping manipulation in general is less than one percent of the way to human level from where it was 20 years ago).</p> |
| </HTML> | </HTML> |
| |
| |
| <HTML> | <HTML> |
| <p>5. <strong>Jeopardy</strong>: AI appears to have taken two or three years to move from lower ‘champion’ level to surpassing world champion level (see <a href="https://www.aaai.org/ojs/index.php/aimagazine/article/view/2303/2165">figure 9</a>; Watson beat Ken Jennings in <a href="http://en.wikipedia.org/wiki/Watson_%28computer%29">2011</a>). We don’t know how far ‘champion’ level is from the level of a beginner, but would be surprised if it were less than four times the distance traversed here, given the situation in other games, suggesting a minimum of a decade for crossing the human spectrum.</p> | <p>8. <strong>Jeopardy</strong>: AI appears to have taken two or three years to move from lower ‘champion’ level to surpassing world champion level (see <a href="https://www.aaai.org/ojs/index.php/aimagazine/article/view/2303/2165">figure 9</a>; Watson beat Ken Jennings in <a href="http://en.wikipedia.org/wiki/Watson_%28computer%29">2011</a>). We don’t know how far ‘champion’ level is from the level of a beginner, but would be surprised if it were less than four times the distance traversed here, given the situation in other games, suggesting a minimum of a decade for crossing the human spectrum.</p> |
| </HTML> | </HTML> |
| |
