Understanding speech recognition of Amazon Alexa and Google Assistant - alexa

According to my experience I have found out following outcomes which may not be correct. I want your opinions on that.
When I design interaction model in Alexa, then Alexa's speech recognition seems to adapt with the change in language model.
In case of Google Assistant, (I am using dialogflow here) any change in language model is not helping with the speech recognition. It seems to use google's first intuition to resolve speech into text.

If you use entities, then those are passed on to the Assistant platform to help with speech biasing.

Related

What are the AI features of Watson Assistant?

It is hard to tell from IBM's documentation but what features of Watson Assistant are actually based on AI?
Features based on NLU (Natural Language Understanding) that they named, also known as NLP(Natural Language Processing) in AI field. Using this service we can implement chatbot or other AI based app.

Can I use my own voice recognition API for Arabic language in Alexa skill?

I already have a voice recognition API for Arabic language but Alexa does not support Arabic language so can I use my API?
No, because at the moment you have to define all your intents and sample utterances in Alexa interaction model (https://developer.amazon.com ). At the moment it supports only English-US, English-UK, and German. As you cannot configure the Arabic Language in that you cannot interact with Alexa in Arabic. Maybe in feature release, Amazon will come up with more language support.

Ai Assistant programming guidance needed

Can any one tell me or guide me in programming an ai Assistant something like Jarvis or Google Assistant etc which has both online and offline voice recognition capability.
I am new to Ai so I tried many tuorials and all still not able to understand or build one. Also don't know where to start and how to start please any help I really need help.
To be frank, natural language processing is one of the most complex and technically difficult fields in computer science. so if you want to make your own version of google assistant, it would help to have an advanced degree in AI, a million dollars in research funding and your own team of engineers.
That being said, a chatbot makes for a really fun hobby project. For now, try not to worry about online and offline voice recognition capability. Make a text-based chatbot that handles basic conversation. You can always add more capability later, and you'll probably have your bearings by then and know what to do.
A good place may be microsoft's new bot framework. I've never used it, myself, but its goal is to take some of the technologies behind the likes of Google Assistant and Jarvis, and to make them available for the everyday developer. It seems to fit your use-case, and as a microsoft product, it'll (probably) have some documentation or tutorials for you to get started.
There are a couple of options to get started.
First off, try to build a
bot using C# for native windows
applications. Microsoft has great documentation for the same, and there are a couple of great tutorials on YouTube for the same.
You can also try
api.ai
to build a bot. It's a bit less hands on, but a good way to get started.
To really try doing everything yourself, try learning a bot of machine learning first. Google has great YouTube tutorials for the same.
Try:
C# bot on windows
Google machine learning
The best choice to start is api.ai. It is simple to learn and integrate, and have a good response time. I tried most of the chatbot engines, apply to the natural language by phone to build voice assistants (Voximal). An important factor in this case is the response time. If you plan to integrate a lot of complex datas the reponse time will increase, and remember that you need to add the time duration of the SpeechToText and TextToSpeech too...
Use my project as an inspiration, is a personal A.I. assistant that is running on Windows 10/11(maybe even 8, not tested). It uses Tokenization and content analysis and association with set parameters for natural language processing and offline and online speech recognition for speech recognition. It can search content on Amazon, Google, Google Images, Google News, Netflix, Wikipedia and eBay. It can open and close multiple applications and it can also navigate in the settings menu on windows on any page or sub-section.
The project is here: https://github.com/CSharpTeoMan911/Eva

Using the robot Pepper with Watson from IBM

I've read of the project in the news, but I can't find the actual state: Is it already possible to use the robot Pepper with Watson (IBM). If yes, what do we need to use Watson with Pepper. Can we do it ourselves? Is there a good tutorial for a getting started with the topic or something similar?
Kind Regards,
Janine
As far as I know, there is nothing else than the classic Watson/Bluemix Rest APIs that you can call from your code. Unfortunately that's all we have right now.
Jonas

What Technologies Is Google Wave Using?

What are the technologies employed in building Google Wave?
Edit: I have moved and amalgamated the answers into an answer below, rather than here in the question where they were.
Real-time editing - some kind of Ajax/Comet for server side calls.
Version-control.
Built in Google Web Toolkit (GWT).
GWT involves Java, Javascript, CSS, HTML.
Custom built protocol. Wave protocol.
Uses the xmmp standard
Rich text-editor.
Language translation.
Google Gears, for drag and drop (of files) functionality. Trying to get it into the HTML5 spec.
It's using HTML5 for the interface and XMPP (Jabber's protocol) for the communication stuff.
The revolutionary part is Operational Transform. Based on the Jupiter collaborative system.
You may like to watch this video: Google I/O 2009 - Google Wave Under the hood
I believe Google Wave is built on HTML5.
I understand that they use an Extension to the Jabber protocol for the federated wave servers.
The realtime editing is based on an algorithm sometimes called jupiter algorithm and described in this paper. More informations can be found on http://www.waveprotocol.org/whitepapers

Resources