AI Zone Admin Forum Add your forum

NEWS: Chatbots.org survey on 3000 US and UK consumers shows it is time for chatbot integration in customer service!read more..

Building intelligent chatbots without voice.. Not possible???
 
 
  [ # 16 ]
Victor Shulist - Aug 17, 2010:

Awesome stuff !!
So it will be the combined strengths of both audio waveform analysis and contextual/knowledge based inference !  Now that is truly fascinating !!.  When I’m completed my core NLP engine, I will want to look into that, putting that ‘front end’ for the bot..

Actually, after setting up the animation forum (3d Humans), a forum dedicated to voice is on my wish list grin.

Nathan Hu - Sep 7, 2010:

I believe it’s not the main problem we are facing. When human chat in text, they get different information from voice chatting. As Erwin’s example,  in all the sentences, the basic idea are the same. The only difference is that besides the basic idea the listener may guess some additional information. In text chatting, the reader won’t guess. Actually, when a human chat in text with another human. He may use two sentences to represent the full information instead of only one sentence with intonation.

I get what you mean. Actually what you’re saying is that chatbot who can communicate without voice must even be more intelligent than voice chatbots, just because they have less information. They can’t use sound to determine the language for example, they only can use the words, and typing speed. So a chatbot in text is more difficult to build than a chatbot in voice, and we always assumed it was the other way around.

@Dave: thanks for coloring my day raspberry (I needed it, it’s raining again while it’s still summer in the northern hemisphere :-s)

 

 
  [ # 17 ]

The problem is that a voice chatbot need to recognize the voice at first.
The accuracy of speech recognition is not good enough still.
Actually, I believe the semantic knowledge is also important for the speech recognition.
The voice communication between human is not easier than text chatting yet.

Erwin Van Lun - Sep 7, 2010:

I get what you mean. Actually what you’re saying is that chatbot who can communicate without voice must even be more intelligent than voice chatbots, just because they have less information. They can’t use sound to determine the language for example, they only can use the words, and typing speed. So a chatbot in text is more difficult to build than a chatbot in voice, and we always assumed it was the other way around.

quote]

 

 
  [ # 18 ]

Voice recognition is further than we generally tend to think. Have you seen this:

http://www.chatbots.org/conversational/agent/cassandra_speech_conversational_agent/

awesome, isn’t it?

 

 
  [ # 19 ]

Many misunderstandings in email - or in a forum - result from things like this. Missing intonation, which leads to a wrong interpretation/disambiguation and voila: Let the flamewars begin. Email is not the only domain where intonation causes problems. See sign language and deaf people. There are concepts for putting something in focus, but it still does not resemble the possibilities of the audio channel.

Very true indeed. Even more, people’s words in combination with their face and body posture often tells so much more (and often contradicting information) compared to just the words.

Another interesting question perhaps is: at what level should you extract this ‘extra’ info from the chatbot’s input data? What I mean is, should it be supplied as ‘tags’ like html, which the system can understand, or do you try to perform a ‘visual’ analysis’ on the text. After all, for us, text is a ‘visual’ thing, not auditory.

 

 
  [ # 20 ]
Jan Bogaerts - Sep 12, 2010:

Another interesting question perhaps is: at what level should you extract this ‘extra’ info from the chatbot’s input data? What I mean is, should it be supplied as ‘tags’ like html, which the system can understand, or do you try to perform a ‘visual’ analysis’ on the text. After all, for us, text is a ‘visual’ thing, not auditory.

I think we shouldn’t try to convert non textual observations into text, just for the sake of our understanding. I personally believe true AI will record its observations, compare it to the model of the world, adjust the model when necessary, respond to it, and adjust the model of the world again. And the model is something we haven’t modelling before, but somehting that grows organically.

 

 < 1 2
2 of 2
 
  login or register to react