Module 1: Introduction to Artificial Intelligence Flashcards
The term Artificial Intelligence was first coined by ___ at the Dartmouth Conference.
John McCarthy
The term Artificial Intelligence was first coined by John McCarthy at the ___.
Dartmouth Conference
The Science and Engineering of making intelligent machines, especially intelligent computer programs.
Artificial Intelligence
AI is the ___ and ___ of making intelligent machines, especially intelligent computer programs.
Science, Engineering
AI is the theory and development of computer systems able to perform task that normally require ___, such as visual perception, speech recognition, decision making, and translation between languages.
human intelligence
AI is a technique of getting machines to work and behave like ___.
humans
It is the computational part of the ability to achieve goals in the world.
intelligence
Varying kinds and degrees of intelligence occur in people, many animals and some machines.
A Rough Classification of AI
- Systems that think like humans
- Systems that act like humans
- Systems that think rationally
- Systems that act rationally
“The art of creating machines that perform functions that require intelligence when performed by people.” (Kurzweil, 1990)
Systems that act like humans
Turing test
Systems that act like humans: Turing Test
“The art of creating machines that perform functions that require intelligence when performed by people.”
- ____
Kurzweil (1990)
“The study of how to make computers do things at which, at the moment, people are better.” (Rich & Knight, 1991)
Systems that act like humans
Turing test
Systems that act like humans: Turing Test
“The study of how to make computers do things at which, at the moment, people are better.”
- ___
Rich and Knight (1991)
You enter a room which has a computer terminal. You have a fixed period of time to type what you want into the terminal, and study the replies. At the other end of the line is either a human being or a computer system.
If it is a computer system, and at the end of the period you cannot reliably determine whether it is a system or a human, then the system is deemed to be intelligent.
Turing Test
The Turing Test approach
* A human questioner cannot tell if there is a computer or a human answering his question, via teletype (remote communication)
* The computer must behave intelligently
Intelligent behavior
* to achieve human-level performance in all cognitive tasks
These cognitive tasks include:
- Natural language processing - for communication with human
- Knowledge representation - to store information effectively & efficiently
- Automated reasoning - to retrieve & answer questions using the stored information
- Machine learning - to adapt to new circumstances
Includes two more issues:
- Computer vision - to perceive objects (seeing)
- Robotics - to move objects (acting)
Systems that act like humans
Turing test
The total Turing Test
- A human questioner cannot tell if there is a computer or a human answering his question via teletype (remote communication)
- The computer must behave intelligently
The Turing Test approach
To achieve human-level performance in all cognitive tasks
Intelligent behavior
Systems that act like humans: Turing test
Cognitive tasks that are included:
* Natural language processing - for communication with human
* Knowledge representation - to store information effectively & efficiently
* Automated reasoning - to retrieve & answer questions using the stored information
* Machine learning - to adapt to new circumstances
Noted
Systems that act like humans: Turing test
Cognitive tasks that are included:
* ___
* ___
* ___
* ___
- Natural language processing - for communication with human
- Knowledge representation - to store information effectively & efficiently
- Automated reasoning - to retrieve & answer questions using the stored information
- Machine learning - to adapt to new circumstances
Systems that act like humans: Turing test
For communication with humans
Natural language processing
Systems that act like humans: Turing test
To store information effectively & efficiently
Knowledge representation
Systems that act like humans: Turing test
To retrieve & answer questions using the stored information
Automated reasoning
Systems that act like humans: Turing test
To adapt to new circumstances
Machine learning
The Total Turing Test
The total Turing Test includes two more issues:
* ___
* ___
- Computer vision - to perceive objects (seeing)
- Robotics - to move objects (acting)
The Total Turing Test Issues
To perceive objects (seeing)
Computer vision
The Total Turing Test Issues
To move objects (acting)
Robotics
Systems that think like humans: Cognitive Modeling
* Humans as observed from ‘inside’
* How do we know how humans think?
* Introspection vs. psychological experiments
* Cognitive Science
* “The exciting new effort to make computers think … machines with minds in the full and literal sense”
(Haugeland, 1985)
* “[The automation of] activities that we associate with human thinking, activities such as decision making, problem solving, learning…” (Bellman, 1978)
Noted
Humans as observed from ‘inside’
Systems that think like humans
Cognitive modeling
How do we know how humans think?
* Introspection vs. psychological experiments
Systems that think like humans
Cognitive modeling
Cognitive Science
Systems that think like humans
Cognitive modeling
“The exciting new effort to make computers think … machines with minds in the full and literal sense” (Haugeland, 1985)
Systems that think like humans
Cognitive modeling
Systems that think like humans: Cognitive Modeling
“The exciting new effort to make computers think … machines with minds in the full and literal sense”
- ___
Haugeland (1985)
“[The automation of] activities that we associate with human thinking, activities such as decision making, problem solving, learning…” (Bellman, 1978)
Systems that think like humans
Cognitive modeling
Systems that think like humans: Cognitive Modeling
“[The automation of] activities that we associate with human thinking, activities such as decision making, problem solving, learning…”
- ___
Bellman (1978)
Systems that think ‘rationally’: “laws of thought”
* Humans are not always ‘rational’
* Rational - defined in terms of logic.
* Logic can’t express everything (e.g. uncertainty)
* Logical approach is often not feasible in terms of computation time (needs ‘guidance’)
* “The study of mental facilities through the use of computational models” (Charniak & McDermott, 1985)
* “The study of the computations that make it possible to perceive, reason, and act” (Winston, 1992)
Noted
“laws of thought”
Systems that think rationally
Systems that think ‘rationally’ use laws of ___.
thought
Defined in terms of logic
Rational
___ cannot express everything.
Logic
“The study of mental facilities through the use of computational models” (Charniak & McDermott, 1985)
Systems that think rationally
Laws of thought
Systems that think ‘rationally’: “laws of thought”
“The study of mental facilities through the use of computational models”
- ____
Charniak and McDermott (1985)
“The study of the computations that make it possible to perceive, reason, and act” (Winston, 1992)
Systems that think rationally
Systems that think ‘rationally’: “laws of thought”
“The study of the computations that make it possible to perceive, reason, and act”
- ____
Winston (1992)
Systems that act rationally: “Rational agent”
* Rational behavior: doing the right thing
* The right thing: that which is expected to maximize goal achievement, given the available information
* Giving answers to questions is ‘acting’.
* “Computational Intelligence is the study of the design of intelligent agents.” (Poole et al., 1998)
* “Al … is concerned with intelligent behavior in artifacts.” (Nilsson, 1998)
Noted
“Rational agent”
Systems that act rationally
Doing the right thing
Rational behavior
That which is expected to maximize goal achievement, given the available information
The ‘right thing’
Systems that act rationally: “Rational agent”
Giving answers to questions is ‘___’.
acting
“Computational Intelligence is the study of the design of intelligent agents.” (Poole et al., 1998)
Systems that act rationally
“Rational agent”
“Computational Intelligence is the study of the design of intelligent agents.”
* ___
Poole et al. (1998)
“Al … is concerned with intelligent behavior in artifacts.” (Nilsson, 1998)
“Rational agent”
“Rational agent”
“Al … is concerned with intelligent behavior in artifacts.”
* ___
Nilsson (1998)
___ was a giant animated bronze warrior who was programmed to guard the island of Crete.
Talos
*700 BC
He circled the island’s shores three times daily.
Talos was a giant animated ___ who was programmed to guard the island of Crete from pirates and invaders.
bronze warrior
Talos was a giant animated bronze warrior who was programmed to guard the island of ___ from pirates and invaders.
Crete
In ___, Alan Turing created the ___. This test is basically used to determine whether or not a computer can think intelligently like a human being.
1950, Turing Test
___ created the Turing Test.
Alan Turing
This test is basically used to determine whether or not a computer can think intelligently like a human being.
Turing Test
Turing introduced the test in his 1950 paper called ___ while at the University of Manchester.
Computing Machinery and Intelligence
In Turing’s paper, Turing proposed a twist on what is called ___. It involves no use of AI, but rather three human participants in three separate rooms. Each room is connected via a screen and keyboard, one containing a male, the other a female, and the other containing a male or female judge. The female tries to convince the judge that she is the male, and the judge tries to disseminate which is which.
The Imitation Game
Turing changes the concept of this game to include an AI, a human and a human questioner. The questioner’s job is then to decide which is the AI and which is the human. Science the formation of the test, many AI have been able to pass; one of the first is a program created by Joseph Weizenbaum called ELIZA.
- Eliza is a mock Rogerian ___.
- The original program was described by Joseph Weizenbaum in 1966.
- This implementation by Norbert Landsteiner 2005.
psychotherapist
Eliza is a mock ___ psychotherapist.
Rogerian
The original Eliza program was described by ___ in ___.
Joseph Weizenbaum, 1966
Eliza was implemented by ___ in ___.
Norbert Landsteiner, 2005
How is the Turing Test used today?
* The original format of the test is still used to this day
* For example, the Loebner Prize has been awarded annually since 1990 to the most human-like computer program as voted by a panel of judges.
* The competition follows the standard rules of the Turing Test. Critics of the award’s relevance often downplay it as more about publicity than truly testing if machines can think.
Noted
The format of the competition was that of a standard Turing test. In each round, a human judge simultaneously holds textual conversations with a computer program and a human being via computer. Based upon the responses, the judge must decide which is which.
The ___ has been awarded annually since ___ to the most human-like computer program as voted by a panel of judges.
Loebner Prize, 1990
The contest was launched in 1990 by Hugh Loebner in conjunction with the Cambridge Center for Behavioral Studies, Massachusetts, United States. Since 2014 it has been organised by the AISB at Bletchley Park. It has also been associated with Flinders University, Dartmouth College, the Science Museum in London, University of Reading and Ulster University, Magee Campus, Derry, UK City of Culture.
In 2004 and 2005, it was held in Loebner’s apartment in New York City. Within the field of artificial intelligence, the Loebner Prize is somewhat controversial; the most prominent critic, Marvin Minsky, called it a publicity stunt that does not help the field along.
In 2019 the format of the competition changed. There was no panel of judges. Instead, the chatbots were judged by the public and there were to be no human competitors.
It is computer program pretending to be a 13-year-old Ukrainian boy that passed a Turing test at the Royal Society in London on the 60th anniversary of Turing’s death by convincing 33% of the judges that it was human during a five-minute typed conversation.
Eugene Goostman
Eugene Goostman was created in ___.
2014
The ___ can respond to calls and book appointments for you.
Google Duplex
With the use of ___ machine, computer scientist Christopher Strachey wrote a checkers program. The first attempt at creating programs that could play chess or that would compete with humans in playing chess.
Ferranti Mark 1
With the use of Ferranti Mark 1 machine, computer scientist ___ wrote a checkers program in 1951.
Christopher Strachey
In ___, Christopher Strachey wrote a ___ program.
1951, checkers
The first attempt at creating programs that could play chess or that would compete with humans in playing chess.
The term Artificial Intelligence was coined by John McCarthy at the Dartmouth Conference in ___.
1956
In ___, the ___ was established. The first AI laboratory which is still present until now.
1959, MIT Lab
In ___, the first ___ was introduced to General Motors assembly line.
1960, robot
In 1960, the first robot was introduced to ___ assembly line.
General Motors
In ___, the first chatbot named ___ was invented.
1966, Eliza
History of chatbots – From ELIZA to Alexa
ELIZA
* was the very first chatbot.
* was created by Joseph Weizenbaum in 1966
* uses pattern matching and substitution methodology to simulate conversation
PARRY
* was constructed by American psychiatrist Kenneth Colby in 1972.
* The program imitated a patient with schizophrenia.
* It attempts to simulate the disease.
* It is a natural language program that resembles the thinking of an individual.
JABBERWACKY
* the chatbot was created by developer Rollo Carpenter in 1988.
* it aimed to simulate a natural human conversation in an entertaining way.
DR. SBAITSO
* is a chabot created by Creative Labs for MS-Dos in 1992.
* is recognized for its full voice operated chat program
* the program would converse with the user as if it was a psychologist
A.L.I.C.E. (Artificial Linguistic Internet Computer Entity)
* is a universal language processing chatbot that uses heuristic pattern matching to carry conversations
* Richard Wallace pioneered the construction of ALICE
* It was formerly known as Alicebot because it was first to run on a computer by the name of Alice
* The program works with the XML schema known as artificial intelligence markup language (AIML)
SmarterChild
* was in many ways a precursor of Siri and was developed in 2001
* was available on AOL, IM, and MSN Messenger with the strength to carry out fun conversations with quick data access to other services
Siri
* was formed by Apple for iOS in 2010
* it is an intelligent personal assistant and learning navigator that uses a natural language UI
Google Now/Google Assistant
* was launched at Google Inch in 2012
* It answers questions, performs actions through requests made to a set of web services and makes recommendations
Cortana
* was first demonstrated at Microsoft’s Build 2014 developer conference
* it became directly integrated into both Windows phone devices and** Windows 10 PCs**
Alexa
* is an intelligent personal assistant developed by Amazon
* It was introduced in 2014
* is now built in to devices such as the Amazon Echo, the Echo Dot, the Echo Show and more
* There is also an Alexa app and more devices from third-party manufacturers that have Alexa built in to them
* All you have to do is say “Alexa, play some music” or “Alexa, find me an Italian restaurant” and she will help you out
Noted
The very first chatbot
ELIZA
The program was designed in a way that it mimics human conversation. The Chatbot ELIZA worked by passing the words that users entered into a computer and then pairing them to a list of possible scripted responses. It uses a script that simulated a psychotherapist. The script proved to be a significant impact on natural language processing and unnatural intelligence, with copies and variants protruding up at academies around the country.
However, Weizenbaum was troubled by the reaction of users. He intended ELIZA to be a mere caricature of human conversation, yet suddenly users were confiding their most profound thoughts in ELIZA. Experts were declaring that chatbots would be indistinguishable from humans within a small number of years.
Weizenbaum rejected the notion that machines could replace human intellect. He argued instead that such devices were just tools, and extensions of the human mind. He further stressed that computers’ understanding of language was entirely dependent on the context in which they were used. Furthermore, Weizenbaum argued that a more general computer understanding of human language was not possible.
In the decades that followed, chatbot makers have built upon Weizenbaum’s model to strive for more human-like interactions. Passing the Turing test has grown to a common goal, which tests new bots’ conversational talents against a board of human judges. The hardest thing in the Turing test issue is that there’s no limit on what people can discuss.
In 1966, Eliza was created by ___.
Joseph Weizenbaum
ELIZA uses ___ and ___ to simulate conversation.
pattern matching,
substitution methodology
It gave users an illusion of understanding on the part of the program, but had no built in framework for contextualizing events.
Directives on how to interact were provided by “scripts”, written originally in MAD-Slip, which allowed ELIZA to process user inputs and engage in discourse following the rules and directions of the script.
The most famous script, DOCTOR, simulated a Rogerian psychotherapist (in particular, Carl Rogers, who was well-known for simply parroting back at patients what they had just said), and used rules, dictated in the script, to respond with non-directional questions to user inputs.
As such, ELIZA was one of the first chatterbots and one of the first programs capable of attempting the Turing test.
In 1972, PARRY was constructed by American psychiatrist ___.
Kenneth Colby
In ___, ___ was constructed by American psychiatrist Kenneth Colby.
1972, PARRY
PARRY imitated a patient with ___. It attempts to simulate the disease.
schizophrenia
PARRY is a ___ program that resembles the thinking of an individual.
natural language
PARRY works via a complicated system of assumptions, attributions, and “emotional responses” triggered by changing weights assigned to verbal inputs. To validate the work, PARRY was tested using a variation of the Turing test. It was in the early seventies when human interrogators, interacting with the program via a remote keyboard, were weak with more than random accuracy to distinguish PARRY from an original paranoid individual.
Fifty years ago, Kenneth Mark Colby was the only psychiatrist thinking about how computers could contribute to the understanding of mental illness. He thus began the project „Overcoming Depression” that lasted until his death in 2001.
JABBERWACKY was created by developer ___ in 1988.
Rollo Carpenter
In ___, ___ was created by developer Rollo Carpenter.
1988, JABBERWACKY
Jabberwacky has led to other technological growth. Some individuals use it for academic research purposes through its webpage since its origin.
JABBERWACKY aimed to simulate a natural human conversation in an ___way.
entertaining
The chatbot is considered to use an AI technique called “contextual pattern matching.”
It is a chabot created by Creative Labs for MS-Dos in 1992.
DR. SBAITSO
It is one of the earliest efforts of incorporating A.I. into a chatbot and is recognized for its full voice operated chat program.
The program would converse with the user as if it was a psychologist. Most of its responses were along the lines of “Why do you feel that way?” rather than any sort of complicated interaction.
In ___, DR. SBAITSO was created by ___ for ___.
1992, Creative Labs, MS-Dos
DR. SBAITSO is recognized for its ___ operated chat program.
full voice
DR. SBAITSO would converse with the user as if it was a ___.
psychologist
It is a universal language processing chatbot that uses heuristic pattern matching to carry conversations.
A.L.I.C.E. (Artificial Linguistic Internet Computer Entity)
The program simulates chatting with a real person over the Internet. Alice is a young-looking woman in human years and tells a user her age, hobbies and other fascinating facts, as well as answering to the user’s dialog.
___ pioneered the construction of A.L.I.C.E.
Richard Wallace
A.L.I.C.E. was formerly known as ___ because it was first to run on a computer by the name of Alice.
Alicebot
A.L.I.C.E. works with the XML schema known as ___.
artificial intelligence markup language (AIML)
The program works with the XML schema known as artificial intelligence markup language (AIML), which helps specify conversation rules. In 1998, the program was edited in Java, and in 2001 Wallace printed an AIML specification. From there, other developers drafted free and open sources of ALICE in different programming languages and a variety of foreign languages.
In ___, a precursor of Siri ___ was developed.
2001, SmarterChild
SmarterChild was available on ___, ___, and ___ Messenger with the strength to carry out fun conversations with quick data access to other services.
AOL, IM, MSN
It suites Microsoft also built its own SmarterChild, years later after most people stopped using AIM which targeted 18- to 24-year-olds in the U.S. the account suites particular conversation.
In ___, Apple formed ___ for iOS.
2010, Siri
A patent application by the United States Patent and Trademark Office details a new Apple service where users could make inquiries and conversation with Siri through Messages. The new patent is similar to a published late last year, but now includes deeper integration with audio, video, and image files.
Similar to other texting and Facebook Messenger Apple’s patent describes a Siri that could perform current duties without the user having to chat aloud. That could be helpful in several public spheres.
They could reply to a text, audio, images, and video when transferred to it by the user. Apple said this would result in more fruitful interactive experience among a consumer and a digital assistant.
The patent provides a few examples of a conversation held between Siri and a user in Messages, with the user asking questions.
It is an intelligent personal assistant and learning navigator that uses a natural language UI.
Siri
In ___, Google Inch launched ___.
2012, Google Now/Google Assistant
It answers questions, performs actions through requests made to a set of web services and makes recommendations.
Google Now/Google Assistant
It was part of a package of updates and UI modifications for mobile search, which included a female-voiced portable assistant to compete with Apple’s Siri.
Google Now was initially a way to get contextually appropriate information based on location and time of the day. It evolved to become much more complicated and elaborate, with a broad range of content categories delivered on cards.
Sometimes it refers to us as predictive search. Currently, it’s built for use in smartphones and has been upgraded to accommodate several features.
Google Now was replaced by Google Assistant in 2017. Today, the assistant is part of a more aggressive Google search growth strategy. The idea is simple, Google wants to provide information in an easy-to-read format before you even know you need it.
It was first demonstrated at Microsoft’s Build 2014 developer conference and became directly integrated into both Windows phone devices and Windows 10 PCs.
Cortana
This program uses voice recognition and relevant algorithms to get and respond to voice commands.
For someone to get started, he or she must type a question in the search box, or select the microphone and talk to Cortana. If a person is not very sure of what to say, he or she will see suggestions on the lock screen, as well as in Cortana home by selecting the search box on the taskbar.
Cortana can perform tasks like reminders based on time, places, or people, send emails and texts, create and manage lists, chit-chat, and play games, find facts, files, locations, and info among others.
Cortana was first demonstrated at Microsoft’s ___ conference and became directly integrated into both Windows phone devices and Windows 10 PCs.
Build 2014 developer
It is an intelligent personal assistant developed by Amazon.
Alexa
- is now built in to devices such as the Amazon Echo, the Echo Dot, the Echo Show and more.
- There is also an Alexa app and more devices from third-party manufacturers that have Alexa built in to them.
- All you have to do is say “Alexa, play some music” or “Alexa, find me an Italian restaurant” and she will help you out.
In ___, Amazon introduced ___.
2014, Alexa
In ___, IBM’s ___ beats the world champion Garry Kasparov in the game of chess.
1997, Deep Blue
In 1997, IBM’s Deep Blue beats the world champion ___ in the game of chess.
Garry Kasparov
In ___, a robotic car named ___, built by Stanford’s racing team, won the DARPA Grand Challenge.
2005, Stanley
In 2005, a robotic car named Stanley, built by ___, won the DARPA Grand Challenge.
Stanford’s racing team
Stanley won the ___ Grand Challenge in 2005.
DARPA (Defense Advanced Research Projects Agency)
In ___, IBM’s question answering system, ___, defeated the two greatest Jeopardy champions, Brad Rutter and Ken Jennings.
2011, Watson
In 2011, IBM’s question answering system, Watson, defeated the two greatest Jeopardy champions, ___ and ___.
Brad Rutter, Ken Jennings
In ___, two researchers Jeff Dean and Andrew Ng from Google trained a _ to recognize cats by showing it unlabeled images and no background information.
2012, neural network
In 2012, two researchers ___ and ___ from Google trained a neural network to recognize cats by showing it unlabeled images and no background information.
Jeff Dean, Andrew Ng
In ___, Musk, Wozniak and Hawking urge ban on ___ and ___.
2015,
warfare AI, autonomous weapons
In 2015, ___, ___, and ___ urge ban on warfare AI and autonomous weapons.
Musk, Wozniak, Hawking
In ___, Hanson Robotics created a humanoid robot named ___, who became known as the first “robot citizen” and was the first robot created with a realistic human appearance and the ability to see and replicate emotions, as well as to communicate.
2016, Sophia
Sophia was created by ___.
Hanson Robotics
The first “robot citizen” and was the first robot created with a realistic human appearance and the ability to see and replicate emotions, as well as to communicate.
Sophia
In ___, ___ was able to converse and learn how to negotiate, but as they went back and forth they ended up forgoing English and developing their own language, completely autonomously.
2017, Facebook AI chatbots
In ___, A Chinese tech group called ___’s language-processing AI beat human intellect on a Stanford reading and comprehension test.
2018, Alibaba
In ___, Google’s ___ reached Grandmaster on the video game StarCraft 2, outperforming all but .2% of human players.
2019, AlphaStar
In 2019, _’s AlphaStar reached Grandmaster on the video game ___ outperforming all but .2% of human players.
Google, StarCraft 2
In 2019, _Google’s AlphaStar reached Grandmaster on the video game StarCraft 2 outperforming all but ___ of human players.
0.2%
In ___, OpenAI started beta testing ___.
2020, GPT-3
In ___, OpenAI developed ___.
2021, DALL-E
It is an AI program that creates images from text descriptions.
Demand for AI
- More computational power
- More data
- Better algorithms
- Broad investment
The first reason is that we have more computational power now. Artificial intelligence requires a lot of computing power. Recently, many advances have been made and complex deep learning models are deployed. And of the greatest technology that made this possible are GPUs.
Second most important reason is that we have a lot of data at present. We’re generating data at an immeasurable pace. We are generating data through social media, through IoT devices. Every possible way, there’s a lot of data. So we need to find a method or a solution that can help us process this much data, and help us derive useful insight, so that we can grow business with the help of data.
Next reason is now we have better algorithms. Right now we have very effective algorithms which are based on the idea of neural networks. Neural networks is nothing but the concept behind deep learning. Since we have better algorithms, which can do better computations and quicker computations with more accuracy, the demand for AI has increased.
Another reason is that universities, governments,startup, and techgiants are all inversting in AI. Companies like Google, Amazon, Facebook, Microsoft, all of these companies have heavily invested in AI because they believe that AI is the future.
Computational Requirements
- If your tasks are small and can fit in a complex sequential processing, you don’t need a big system. You could even skip the GPUs altogether.
- A CPU such as i7–7500U can train an average of ~115 examples/second. So, if you are planning to work on other ML areas or algorithms, a GPU is not necessary.
- If your task is a bit intensive, and has a manageable data, a reasonably powerful GPU would be a better choice for you. A laptop with a dedicated graphics card of high end should do the work.
- There are a few high end (and expectedly heavy) laptops like Nvidia GTX 1080 (8 GB VRAM), which can train an average of ~14k examples/second.
- In addition, you can build your own PC with a reasonable CPU and a powerful GPU, but keep in mind that the CPU must not bottleneck the GPU. For instance, an i7-7500U will work flawlessly with a GTX 1080 GPU.
- If you are working on complex problems or are a company that leverages deep learning, you should probably build your own deep learning system or use a cloud service.
- If your task is of a larger scale than usual, and you have enough money to cover up the cost, you can opt for a GPU cluster and do multi-GPU computing. Also, there are more powerful options available – TPUs and faster FPGAs – which are designed specifically for these purposes.
- A GPU can perform convolutional/CNN or recurrent neural networks/RNN based operations. It can also perform operations on a batch of images of 128 or 256 images at once in just a few milliseconds. However, the power consumption is around ~250W and requires a full PC that additionally requires 150W of power, which leads to a total of 400W.
- Applications like virtual or augmented reality goggles, drones, mobile devices, and small robots do not have this much power. Also, in case of autonomous cars and smart cameras, where live video is necessary, image batching is not possible, as video has to be processed in real-time for timely responses.
- In the future, we might see more powerful devices that won’t require that much power and work on mobile platforms/devices.
Noted
A CPU such as ___ can train an average of ~___ examples/second. So, if you are planning to work on other ML areas or algorithms, a GPU is not necessary.
i7–7500U, 115
Laptops like ___ (8 GB VRAM), which can train an average of ~___ examples/second.
Nvidia GTX 1080, 14k
___ are Google’s custom-developed application-specific integrated circuits (ASICs) used to accelerate machine learning workloads.
Tensor Processing Units (TPUs)
It is an integrated circuit that can be programmed by a user for a specific use after it has been manufactured.
Field Programmable Gate Array (FPGA)
AI Applications
- Google Predictive Search Engine
- JPMorgan Chase’s Contract Intelligence (Coin)
- IBM AI Watson technology for medical diagnosis
- Google’s AI Eye Doctor
- Facebook uses Machine Learning & Deep Learning to detect facial features and tag your friends.
- Twitter’s AI is being used to identify any sort of hate speech and terroristic languages in tweets.
- Google’s virtual assistant called the Google Duplex can respond to calls and book appointments and adds human touch.
- TESLA’s self-driving cars
- Netflix Movie Recommendation
- Spam Filtering
JPMorgan Chase’s ___ platform uses Al, machine learning and image recognition software to analyze legal documents.
Contract Intelligence (COiN)
Manually reviewing around 12,000 agreements took over 36,000 hours, that’s a lot of time. But as soon as this task was replaced by AI machine, it was able to do this in a matter of seconds.
___’s Contract Intelligence (COiN) platform uses Al, machine learning and image recognition software to analyze legal documents.
JPMorgan Chase
Healthcare organizations use ___ technology for medical diagnosis.
IBM Al (Watson)
More than 230 healthcare organizations use IBM technology, which is basically IBM Watson.
In 2016, IBM Watson was able to cross reference 20 million oncology records quickly and correctly diagnose a rare leukemia condition in a patient.
So it basically went through 20 million records, which it probably did in a matter of second per minutes, And then it correctly diagnosed a patient with a rare leukemia.
___ can examine retina scans and identify a condition called diabetic retinopathy.
Google’s Al Eye Doctor
The Google’s AI Eye doctor is another initiative, which is taken by Google, where they’re working with an Indian eyecare chain to develop artificial intelligence system which can examine retinal scans and identify a condition called diabetic retinopathy which can cause blindness.
Google’s Al Eye Doctor can examine retina scans and identify a condition called ___.
diabetic retinopathy
___ uses Machine Learning & Deep Learning to detect facial features and tag your friends.
All the auto tagging feature that you see in FB, behind that there’s machine learning, deep learning, neural networks.
___ is being used to identify any sort of hate speech and terroristic languages in tweets.
Twitter’s AI
Recently, the company discovered around 300,000 terroristic link accounts and 95% of these were found by non-human artificially intelligent machines.
Google’s virtual assistant called the ___. It can respond to calls and book appointments while adding human touch.
Google Duplex
In 2018, Google Duplex successfully made an appointment with a ___ over the phone in front of a crowd of ___.
hairdresser, 7000
The receptionist was completely unaware that they weren’t conversing with a real human. This is considered by some to be a modern-day Turing Test pass, despite not relying on the true format of the test as Alan Turing designed it.
More than two years after it initially began trials, Google’s AI-powered reservation service Duplex is now available in 49 US states. This looks like it’ll be the limit of Duplex’s coverage in the US for the time being, as Google tells The Verge it has no timeline to launch the service in the last hold-out state — Louisiana — due to unspecified local laws.
Google wowed audiences when it first unveiled Duplex at its 2018 I/O conference. As a feature of Google Assistant, Duplex uses AI to call local businesses, making reservations at restaurants and hairdressers on your behalf using a realistic-sounding artificial voice.
Initially, it seemed Google promised more than it could deliver. In 2019 it was revealed that 25% of Duplex calls are made by humans, and that 19% of calls started by the automated system have to be completed by people. And in our own reporting, we found that restaurants often confused Duplex with automated spam robocalls. As of October last year (?), though, Google says 99% of Duplex calls are fully automated.
Over ___% of what you watch is recommended by Netflix.
75
Recommendations are made by machine learning.
Types of AI
- Artificial Narrow Language
- Artificial General Language
- Artificial Super Language
Also known as weak AI, it involves applying AI only to specific task.
Artificial Narrow Intelligence (ANI)
Examples include spam filters, voice assistants, and recommendation systems.
Also known as strong AI, it involves machines that possess the ability to perform any intellectual task that a human being can.
Artificial General Intelligence (AGI)
AGI is a major goal of AI research but has not yet been achieved.
It is a term referring to the time when the capability of computers will surpass humans.
Artificial Super Intelligence (ASI)
ASI is purely theoretical and raises many ethical and existential questions.
AI History Summary
* 700 BC* - Talos; Greek Mythology
* 1950 - Turing Test; Alan Turing
* 1951 - Checkers program; Christopher Strachey
* 1956 - Term AI coined; John McCarthy
* 1959 - MIT Lab
* 1960 - First robot; General Motors
* 1966 - Eliza; Joseph Weizenbaum
* 1972 - PARRY; Kenneth Colby
* 1988 - JABBERWACKY; Rollo Carpenter
* 1992 - DR. SBAITSO; Creative Labs
* 1995 - ALICE; Richard Wallace
* 1997 - Deep Blue; IBM
* 2001 - SmarterChild
* 2005 - Stanley; Stanford’s racing team
* 2010 - Siri; Apple
* 2011 - Watson; IBM
* 2012 - Google Now/Google Assistant
* 2012 - Google trained neural network using unlabled images; Jeff Dean & Andrew Ng
* 2014 - Cortana; Microsoft’s Build 2014 developer conference
* 2014 - Eugene Goostman
* 2014 - Alexa; Amazon
* 2015 - Warfare AI and autonomous weapons urge ban; Musk, Wozniak & Hawking
* 2016 - Sophia; Hanson Robotics
* 2017 - Two AI Chatbots developing own language; Facebook
* 2018 - AI winning Stanford reading and comprehension test; Alibaba
* 2019 - AlphaStar; Google
* 2020 - GPT-3 beta testing; OpenAI
* 2021 - DALL-E; OpenAI
Noted
