Marcus L Endicott: Best Quora Answers 2013


See also:

Marcus L Endicott: Best Quora Answers 2011 | Marcus L Endicott: Best Quora Answers 2012 | Marcus L Endicott: Best Quora Answers 2014 | Marcus L Endicott: Best Quora Answers 2015 | Marcus L Endicott: Best Quora Answers 2016


Philosophy of Everyday Life: Will computers ever be able to think “out of the box”?

There is a whole field of study dedicated to Computational creativity.

There is a Quora topic for Computational Dreaming.

To see some of my own answers Search – “computational-dreaming” – Quora:
· Is it possible to explain metaphors using discourse analysis frameworks, or only with cognitive linguistic frameworks?
· What algorithms model the imagination in agent or multi-agent systems?
· How are metaphors handled in AI / NLP / ML?
· Is there a natural language processor in existence that can actually extract semantic content from real human conversation?
· What are examples of “Dream Machines” in Art and Literature?

Written 23 Nov, 2013.

= = =

What does Ifttt think of adding a HTTP request channel?

See the recent reply by Geek Sniper “aka CEO” @Mashape Augusto Marietti to my question: Why doesn’t Mashape make its APIs compatible with IFTTT?

Written 21 Nov, 2013.

= = =

Gmail: Is there any API that can provide flight booking information associated with gmail account?

Both Evature and Skyphrase claim to do this not only for gmail but any email.

See also the recent thread on the Chatbots.org “AI Zone” forum, Email Interfaces to Chatbots.

Written 16 Nov, 2013.

= = =

Survey Questions: What features are you missing in RSS feed readers?

Input, output, and lots and lots of filters…. Google Reader was good, especially with Google Alerts *feeds*. I never understood why Feedburner wasn’t combined with Google Reader. If Goggle Reader would have bought Yahoo Pipes, they could have had the coolest feed reader imaginable…. The main points are feed aggregation, with feed processing, and then feed output. (Hehe, forget about “reading”, which will have a limited shelf life….) Anyway, my dream for Google Reader was always to have the *guts* of Yahoo Pipes…. Oh, and the output should definitely be IFTTT compatible. ;^)

Written 16 Nov, 2013.

= = =

Artificial Intelligence: Quora Feature: Is it possible to build a AI agent that asks/answer questions based on current state of affairs rather than humans posting questions/answers?

Today, most “conversational agent” (aka chatbot) knowledgebases are either hand tooled, for instance by domain experts, or “crowd sourced” by recycling human responses. There are no ready made or turnkey engines available that allow for dynamically modifying the knowledgebase in realtime, for instance via backend API. The problem with “big data” is that it is by very definition too big to work properly in any conventional database system or hardware platform…. See my recent blog post Dissecting the Summarization Processfor some of the nitty gritty involved in doing just this. Otherwise, read through my Quora answers and questions for a veritable primer at Marcus L Endicott.

Written 14 Nov, 2013.

= = =

Bots: Which Twitter bots retweet and/or mention you when prompted or based on specific trigger?

I maintained a list of 100 Best Twitter Bots through 2012; however, since the final Twitter “OAuth apocalypse” with the shutdown of API v1.0 in June 2013 all of my own 250x Meta-Guide Feed Bots have stopped working. Not only that, but my subsequent fix Tweet-2-RSS has been shut down; because, Twitter is now taking a hard line against all non-Twitter verbatim retweeting. I’m sure there are still some responder bots (aka reply bots) using OAuth; but, Twitter also considers this a form of automated spam.

In the early years Twitter was extremely fertile ground for creativity. In fact the original Twitter was sheer brilliance, even if initially mystifying. It is tremendously unfortunate that Twitter was not a public infrastructure like Usenet. The paranoic “Facebook complex” of the Twitter gods and consequent heavy-handed protectionism is guaranteed to turn Twitter into the next Myspace….

Written 12 Nov, 2013. Asked to answer by Mikael S. Pittam.

= = =

Artificial Intelligence: What is the difference and similarity of artificial intelligence with IR?

There is a Quora topic for Information Retrieval. Perhaps one might better ask what the difference and similarity was between information retrieval andsearch, and what the role of search in artificial intelligence may be. As has been pointed out over and over on Quora, artificial intelligence is not a very useful term, but rather a grab bag of disparate sub-domains. For instance, information retrieval is listed under natural language processing on Wikipedia, and NLP may or may not represent a sub-domain of AI. The real problem with “artificial intelligence” is that no one has yet figured out just what “natural intelligence” is….

References:
· Graph-based Natural Language Processing and Information Retrieval
· Learning to Rank for Information Retrieval
· Learning to Rank for Information Retrieval and Natural Language Processing

Written 5 Nov, 2013.

= = =

How does IBM Watson advance question answering in general?

· Yes, IBM Watson is a triumph of brute force computing.
· Yes, Watson is the most advanced integration of existing approaches.
· Apparently, the so-called “ranking” or predictive analytics was the core advancement.

Written 4 Nov, 2013.

= = =

Natural Language Processing: Natural Language Processing for rookies: What are some good references from a Product Manager’s perspective?

It seems to me that a product manager should be first and foremost concerned with competitive intelligence. If you really want the “NLP for rookies” quickstart on what’s out there and what they do, then I strongly suggest you spend some time on my Meta Guide Videography.

Updated 2 Apr.

= = =

What are Augmented Transition Network Grammars?

Let me google that for you… Augmented transition network.

See also my quick and dirty webpage:
· Augmented Transition Network & Dialog Systems – Meta-Guide.com

Written 4 Nov, 2013.

= = =

Natural Language Processing: What are some good papers to read about Question Answering?

Check out the 2011 book:
· Interactive Multi-modal Question-Answering

See also my quick and dirty webpages:
· DBpedia & Question Answering Systems – Meta-Guide.com
· Interactive Question Answering – Meta-Guide.com
· Pythia Question Answering System – Meta-Guide.com
· Question Generators – Meta-Guide.com
· Question Classifier Module – Meta-Guide.com
· Question Processing Module – Meta-Guide.com
· Question Semantic Representation (QSR) – Meta-Guide.com
· Question Answering Module – Meta-Guide.com
· Question Generator Module – Meta-Guide.com
· Question-Answer Pairs & Dialog Systems – Meta-Guide.com
· Question Generation – Meta-Guide.com
· Question Analysis Module – Meta-Guide.com
· SPARQL & Natural Language Question Answering 2011 – Meta-Guide.com
· SPARQL & Natural Language Question Answering 2012 – Meta-Guide.com
· YourQA (York Question Answering System) – Meta-Guide.com

Written 4 Nov, 2013.

= = =

Artificial Intelligence: Which chatbot authoring languages are currently the most flexible and full-featured?

See my previous answer about Bruce Wilcox’s ChatScript:
· What traits should the ideal chatbot authoring language have?

Written 4 Nov, 2013.

= = =

Monetization: How might Google monetize Google Alerts?

I really, really, really want the Google Alerts feed option back that was removed when Google Reader was killed. I would be HAPPY to pay for this in a Google Alerts API.

Written 4 Nov, 2013.

= = =

Audiobooks: Are there any kind of audiobooks that come with text so it can move synced so I can see text if I don’t know a word, or any app that can TTS (Text To Speech) in sync iOS or Android?

Listen to news, websites and documents with SoundGecko… does not “sync” per se, but does allow you to read along while listening.

Updated 6 Nov, 2013.

= = =

Conversational Agents (Chatbots): What’s best way to train an chatbot (e.g. alicebot)?

A “chat log” and AIML file are two different things. Since a chat log should contain both questions and answers, you will have question-answer pairs, which should be enough for statistical/probabilistic analysis. For instance, the University of Southern California (USC) Institute for Creative Technologies (ICT) “Virtual Human Toolkit” (vhtoolkit.ict.usc.edu) includes “NPCEditor”, a tool for building question-answering characters based on statistical language modeling (probabilistic relevance models). An AIML file is a pattern matching system, and I know of no work done applying machine learning to AIML per se; however, chat logs are produced by chatbots such as Alicebot.

Written 4 Nov, 2013.

= = =

Application Programming Interfaces (API): API part of speech parser online?

In the absence of “Question Details”… I’m going to assume you’re referring to a “POS tagger” for Part Of Speech Tagging (POS tagging). There are an increasing number of natural language processing APIs offered on Mashape, such as Stremor Noun Phrase And Part Of Speech Tagging ALPHA API.

Written 1 Nov, 2013.

= = =

Artificial Intelligence: What is the best way to learn and write a AI Chat bot?

The AI Zone forum at Chatbots.org is a happening place.

There are Quora topics for:
· Conversational Agents (Chatbots)
· Dialog System (software)
· Virtual Assistants (bots)

See also my own webpage:
· Theses in AI & NLP (Conversational Agents) – Meta-Guide.com

Written 1 Nov, 2013.

= = =

Quora Usage Data and Analysis: Will quora lend its answer database to an Pseudo-Intelligent Life form, If one is built in the future?

This is becoming a FAQ. See my recent answer to:
· Would an AI be allowed to answer questions on Quora?

Written 30 Oct, 2013.

= = =

Artificial Intelligence: What are some good ways to start on a long AI project?

Jan Bogaerts has a “Neural Network Designer” (Download) for making dialog systems (chatbots). I haven’t tried it, so can’t say how well it works.

The University of Southern California (USC) Institute for Creative Technologies (ICT) “Virtual Human Toolkit” (vhtoolkit.ict.usc.edu) includes “NPCEditor”, a tool for building question-answering characters based on statistical language modeling (probabilistic relevance models).

See the Quora topics:
· Conversational Agents (Chatbots)
· Dialog System (software)
· Virtual Assistants (bots)

See also my quick and dirty webpages:
· NPCEditor – Meta-Guide.com
· Virtual Human Toolkit – Meta-Guide.com

Updated 4 Nov, 2013.

= = =

Machine Learning: Which are top companies in India who expertise in Machine learning, NLP, Unsupervised learning?

See the Quora topic Artificial Intelligence in India. I will be meeting with conversational agent company Dexetra today in Bangalore. (If anyone else in this field would like to meet up in Bangalore during November 2013, message ME: Marcus Endicott.)

Written 28 Oct, 2013.

= = =

Cloud Robotics: What are the communication challenges on cloud robotics that can be worked on?

I have found no good cloud “middleware” for artificial intelligence applications. My ideal would be something visual like Yahoo! Pipes. The main point is how to integrate the proliferating APIs in the cloud, in other words from the browser, and ideally language independent.

There are plug and play SaaS and iPaaS solutions for business processes, but are not adequate for most natural language applications. I’m thinking about something IFTTT simple. In particular, see my 2012 videos from Chatbots 3.2 – Philadelphia.

Written 26 Oct, 2013.

= = =

Artificial Intelligence: Is it possible to write a software that act as a personal assistance?

See the Quora topic Virtual Assistants (bots).

See also my recent answers to:
· Are there any Meta-Assistant bots that utilize or manage multiple assistants with different capabilities?
· What are the technical requirements for building system similar to siri other than mobile app development?

Written 26 Oct, 2013.

= = =

What is the state of the art technology in crawling?

Mozenda is an interesting SaaS that allows you to build “agents” for scraping and crawling.

See also my quick and dirty webpages:
· 100 Best Web Crawler Videos – Meta-Guide.com
· Best Mozenda Videos – Meta-Guide.com

Written 26 Oct, 2013.

= = =

Artificial Intelligence: What’s the techonology behind Korean chat bot simsimi?

Generally speaking, SimSimi appears to operate similarly to Cleverbot by leveraging collective intelligence – in other words recycling the replies it receives. I noticed recently that the SimSimi Conversation API is available on Mashape: SimSimi Conversation API | Mashape. For more information, check the official developer documentation: SimSimi API. ‘Simsimi Ent‘ seems to be on Quora.

See also my quick and dirty webpages:
· 100 Best SimSimi Videos – Meta-Guide.com
· Chatbots Sentiment Analysis: SimSimi vs. Cleverbot – Meta-Guide.com

Written 18 Oct, 2013.

= = =

Siri (software): What are the technical requirements for building system similar to siri other than mobile app development?

See my answers to:
· Which topics on Quora are related to the technology behind Apple Siri?
· Is anyone working on an open source version of Siri?

See also my quick and dirty webpage:
· 500 Best Siri Answers – Meta-Guide.com

Written 17 Oct, 2013.

= = =

International Travel: How would you advise a person who would like to explore the world, yet has no money at all?

Local exchange trading systems (LETS) are something like an alternative to money, on the local level.
· Travelling the world on LETS (ebook by James Taris) details “4 overseas tours, staying FREE with 63 LETS hosts in 19 countries”.

Wikipedia has a Category:Pedestrian circumnavigators of the globe, including:
· Dave Kunst from Minnesota is the first person verified to have walked around the earth; see his 1979 book, The Man Who Walked Around The World.
· Steven Newman from Ohio is listed in the Guinness Book of World Records as the first man to walk solo around the world; see his 1989 book, Worldwalk.
· Ffyona Campbell is a British long-distance walker who was the first woman to walk around the world; see her 1997 book, The Whole Story: A Walk Around The World.

Written 16 Oct, 2013.

= = =

XML to Excel: How can I convert an XML file to an organized Excel output?

XML to CSV via XSLT (Extensible Stylesheet Language Transformations) ..

See my quick and dirty webpage:
· 100 Best XSLT Videos – Meta-Guide.com

Written 15 Oct, 2013.

= = =

Search Engines: What semantic search engines are there?

According to Wikipedia, Semantic search: “Rather than using ranking algorithms such as Google’s PageRank to predict relevancy, semantic search uses semantics, or the science of meaning in language, to produce highly relevant search results.”

See also Wikipedia, List of semantic search engines

Written 14 Oct, 2013.

= = =

What’s a useful API service to do?

A GUI is a graphical user interface for humans. An API is an application programming interface for machines. Ideally, a machine API ought to be able to access all of the human GUI functions of your site or service.

For API ideas, see my answer to:
Need to build a summarizer! That’s okay. But what are the current sub problems that I could address in Summarization? Or what would you like a summarizer do for you? Thinking of a kick ass problem statement.

Written 14 Oct, 2013.

= = =

Cloud APIs: Is there a software to manage Web APIs?

It sounds to me like you need Mashape, otherwise check Apigee.

Written 14 Oct, 2013.

= = =

What IVR does american airlines use?

Apparently, American Airlines was still using Tellme Networks after it was acquired by Microsoft in 2007 (Page on Microsoft). In fact, the domain name tellme[dot]com is currently, incorrectly forwarding to Bing Developer Center: Speech. However, according to Wikipedia, Microsoft divested itsTellme Networks interactive voice response service to 24 7 inc (Intuitive Customer Experience – Mobile, Speech, Web) in early 2012. It is not clear whether American Airlines is currently using Microsoft Tellme, 24 7 inc, or something else.

Written 12 Oct, 2013.

= = =

Bots: Are there any Meta-Assistant bots that utilize or manage multiple assistants with different capabilities?

If you mean is there a mama Siri to manage all the little Siris, the smart home and the smart car, then the answer is no. However, there is a huge amount of research going into Multi-Agent Systems. There are also a number of macrotype programs, that allow you to control other programs, such as AutoHotkey [1], AutoIt [2], Sikuli [3], Selenium [4], Ubot Studio [5], and WinAutomation [6], which could potentially be used for this purpose. And on the dark side, classic Botnets could be described as a form of “Meta-Assistant bots”.

[1] 100 Best AutoHotkey Videos – Meta-Guide.com

[2] 100 Best AutoIt Tutorial Videos – Meta-Guide.com

[3] 100 Best Sikuli Videos – Meta-Guide.com

[4] Best Selenium Videos – Meta-Guide.com

[5] Best Ubot Studio Videos – Meta-Guide.com

[6] Best WinAutomation Videos – Meta-Guide.com

Written 12 Oct, 2013. Asked to answer by Guy Taylor.

= = =

Which application is truly missing in Natural Language Processing?

As nearly as I can tell, most NLP seems to be about tagging. There is precious little about what to actually do with the tags, in other words interpretation. IMHO, what is needed are more NLP interpreters. (In case I’m missing something painfully obvious here, please clue me in.)

Written 12 Oct, 2013.

= = =

Natural Language Processing: What is the state of the art in natural language understanding?

I would say IBM Watson is the leader of the pack. However, Watson was not made in a vacuum; there was quite a bit of other research, not to mention government sponsorship, leading up to it. Of course, Watson was a triumph of brute force computing, which means to me the challenge now is how to make it more economical and accessible to everyman. Also, Watson clearly lacked voice input, so big opportunities for Nuance. Two of the core technologies leveraged by Watson were machine learning and hypothesis ranking.

See my 2011 blog post:
· How Many PlayStations Make A Watson?

Written 12 Oct, 2013.

= = =

Computational Linguistics: What is the difference between a language model and a hierarchical language model?

I’m not a formally trained expert, more like a hobbyist or maven, but a quick search of my Meta-Guide.com website seems to indicate that language models are generally statistical, whereas hierarchical language models are rule or tree based. So, a more precise distinction may be statistical language modeling versus hierarchical language modeling.

See my quick and dirty webpages:
· IRSTLM (IRST Language Modeling) Toolkit – Meta-Guide.com
· Language Modeling & Dialog Systems 2011 – Meta-Guide.com
· Rule-based Language Modeling – Meta-Guide.com
· SRILM (SRI Language Modeling Toolkit) 2011 – Meta-Guide.com

Written 12 Oct, 2013.

= = =

How does NELL work?

1 upvote by Vinit Agrawal.
August 2013 Never-Ending Learning to Read the Web, TheOfficialACM (1:02:11):

Presented by Tom M. Mitchell, Founder and Chair of Carnegie Mellon University’s Machine Learning Department

See also: NELL (cmunell) on Twitter

Written 11 Oct, 2013.

= = =

What algorithms model the imagination in agent or multi-agent systems?

I can’t say specifically what algorithms, though would also like to know. I can say that imagination in agent or multi-agent systems cuts two ways. There is a whole field dedicated to Computational creativity; however, I findComputational Dreaming [1], Daydreaming Machines [2], and Robot Dreams [3] more intriguing. See my recent answer to: How are metaphors handled in AI / NLP / ML?

[1] Computational Dreaming (Draft) | Meta-Guide.com

[2] Daydreaming Machines | Meta-Guide.com

[3] Robot Dreams | Meta-Guide.com

Updated 4 Aug.

= = =

What is middleware or mid-range applications?

Let me google that for you… Middleware. Midrange computer seems to be used in reference to Middleware (distributed applications), primarily for server technology, such as IBM WebSphere in particular on former IBM System i and current IBM i minicomputers.

Written 11 Oct, 2013.

= = =

Text Analytics: Has someone built a conversation guide using text analytics – in a specific context?

1 upvote by Anonymous.
AFAIK, the answer is no. This is a known issue with conversational agents. Generally speaking, all of today’s chatbots are reactive, rather than proactive. Of course, there has been some research done in this area, but with mixed results at best. The big data trend toward personalization technologies does promise progress; generally, some kind of prior knowledge, or sensory information, is necessary for effective proactive communication.

Written 11 Oct, 2013.

= = =

Interactive Voice Response: What is Interactive Voice Response?

Let me google that for you… Interactive voice response. The real difference between Interactive Voice Response and Speech Recognition (AutomaticSpeech Recognition) is that IVR is primarily based on selective, pre-programmed pattern recognition, actually something similar to AIML chatbots; whereas, ASR theoretically works on a higher level, recognizing more sounds and more phrases, if not all. Other than in a practical sense (IVR does less, ASR does more), I’m not convinced the distinction is really valid; I strongly suspect a clever person could program IVR to work better than ASR programmed by a dull person.

Written 11 Oct, 2013.

= = =

When will Quora integrate with AI simulation?

This is becoming a FAQ. See my recent answer to:
· Would an AI be allowed to answer questions on Quora?

Written 11 Oct, 2013.

= = =

Q&A Websites: Are there any web sites that integrate social question-answering with conversational bots?

I would call this the mechanical turk model of crowd sourced, collective intelligence, and I’ve been working on something like it for quite some time at Meta-Guide.com. In particular, check out my video presentation from Chatbots 3.2 – Philadelphia. Typically, question answering systems are based on question answer pairs. Question answer pairs may be crowd sourced, and then form the foundation for a conversational agent knowledgebase.

I’m actually hopping mad at Twitter for effectively closing their API with OAuth; because, I was very close to a non-stop question answering system, sourcing both questions AND answers on Twitter itself…. I’m firmly convinced that the answers to most questions on Twitter have already been posted to Twitter, historically speaking. In fact, most questions about most things have already been posted to Twitter, historically speaking.

There is a whole class of commercial SMS “bots” or agents, that may really be mechanical turks on the back end, probably using crowd sourced replies, offering to provide quick answers to virtually any question for a price, such as Ask Bongo (Q&A system).

Written 9 Oct, 2013.

= = =

Quora Answer Policies and Guidelines: Would an AI be allowed to answer questions on Quora?

There is a topic for Quora Bots (Quora feature). I messaged Quora founderCharlie Cheever in 2011 suggesting… a voluntary, opt-in “bot-zone” as a sandbox for question answering bot projects. To a large extent working with question answering bots would depend on the as yet non-existent Quora API. However, question answering bots may also conflict with the work of Quora founder Adam D’Angelo on his Vicarious (company)… “building software that thinks and learns like a human”.

Written 9 Oct, 2013.

= = =

What’s a good* problem to hack at as a personal project in the area of summarization?

See my answer to:
· Need to build a summarizer! That’s okay. But what are the current sub problems that I could address in Summarization? Or what would you like a summarizer do for you? Thinking of a kick ass problem statement.

Written 8 Oct, 2013.

= = =

Artificial Intelligence: How are metaphors handled in AI / NLP / ML?

See my answers to:
· Is it possible to explain metaphors using discourse analysis frameworks, or only with cognitive linguistic frameworks?
· Where are the interesting topics which intersect Neuroscience and Machine Learning?
· Which professors are doing cutting edge work on Chinese natural language processing?

See also my quick and dirty web pages:
· ACD (Associative Concept Dictionary) – Meta-Guide.com
· ATT-Meta Project – Meta-Guide.com
· Automatic Metaphor – Meta-Guide.com
· Computational Dreaming (Draft) – Meta-Guide.com
· Computational Metaphorics (Draft) – Meta-Guide.com
· CorMet (Metaphor Extraction System) – Meta-Guide.com
· Daydreaming Machines – Meta-Guide.com
· Hamburg Metaphor Database – Meta-Guide.com
· Master Metaphor List – Meta-Guide.com
· MetaBank – Meta-Guide.com
· Metalude – Meta-Guide.com
· Metaphor & Cognition – Meta-Guide.com
· Metaphor Analysis – Meta-Guide.com
· Metaphor Extraction – Meta-Guide.com
· Metaphor Generation 2010 – Meta-Guide.com
· Metaphor Network – Meta-Guide.com
· Metaphor Project – Meta-Guide.com
· Metaphor-based Reasoning – Meta-Guide.com
· MetaViz – Meta-Guide.com
· MetNet – Meta-Guide.com
· MIP (Metaphor Identification Procedure) – Meta-Guide.com
· Robot Dreams – Meta-Guide.com
· SemEval & Metaphors – Meta-Guide.com
· ZMET (Zaltman Metaphor Elicitation Technique) – Meta-Guide.com

Written 8 Oct, 2013.

= = =

Computational Linguistics: I have developed a term extraction program, see http://www.stringmatch.com. How best to reach the target audience?

Personally, I would MUCH prefer to use something like this as a cloud based API, available through Mashape for instance. I could see a hybrid service with a web GUI for certain functions and a commercial API for others.

Written 25 Sep, 2013.

(Visited 1 times, 1 visits today)