Applying pipeline “tensorflow_embedding” of Rasa NLU

According to this nice article, there was a new pipeline released using a different approach from the standard one (spacy_sklearn). I wanted to give it a try to see whether it can help with improving bot’s accuracy.

After applying done, I gave an evaluation of “tensorflow_embedding”. It seemed to work better a bit. For example, I defined intents “greet” and “goodbye” with some following messages in my training data.


## intent:greet
- Hey! How are you?
- Hi! How can I help you?
- Good to see you!
- Nice to see you!
- Hi
- Hello
- Hi there

## intent:goodbye
- Bye
- Bye Bye
- See you later
- Take care
- Peace

In order to play around with Rasa NLU, I created a project here. You can have a look at this change from this pull request. Yay!

When I entered message “hi bot”, then bot with “tensorflow_embedding” could detect intent “greet” with better confidence scores rather than bot with “spacy_sklearn”. The following are responses after executing curl -X POST localhost:5000/parse -d '{"q": "Hi bot", "project": "ctraubot", "model": "nlu"}' | python -m json.tool
A response of bot with pipeline spacy_sklearn

{

"intent": {

"name": "greet",

"confidence": 0.8687479112327237

},

"entities": [],

"intent_ranking": [

{

"name": "greet",

"confidence": 0.8687479112327237

},

{

"name": "goodbye",

"confidence": 0.13125208876727643

}

],

"text": "Hi bot",

"project": "ctraubot",

"model": "nlu"

}

A response of bot with pipeline tensorflow_embedding

{

"intent": {

"name": "greet",

"confidence": 0.9777982831001282

},

"entities": [],

"intent_ranking": [

{

"name": "greet",

"confidence": 0.9777982831001282

},

{

"name": "goodbye",

"confidence": -0.08299092948436737

}

],

"text": "Hi bot",

"project": "ctraubot",

"model": "nlu"

}

By the way, I still need modules spacy and sklearn for a further usage of entity extraction.

Comments

How I did customize "rasa-nlu-trainer" as my own tool

Background I wanted to have a tool for human beings to classify intents and extract entities of texts which were obtained from a raw dataset such as Rocket.chat's conversation, Maluuba Frames or here. Then, the output (labeled texts) could be consumed by an NLU tool such as Rasa NLU.

rasa-nlu-trainer was a potential one which I didn't need to build an app from scratch. However, I needed to add more of my own features to fulfill my needs. They were:

1. Loading/displaying raw texts stored by a database such as MongoDB
2. Manually labeling intents and entities for the loaded texts
3. Persisting labeled texts into the database

I firstly did look up what rasa-nlu-trainer's technologies were used in order to see how to implement my mentioned features.
At first glancerasa-nlu-trainer was bootstrapped with Create React App. Create React App is a tool to create a React app with no build configuration, as it said. This tool is also recommended by the official React.js tutorial. I ac…

AngularJS - Build a custom validation directive for using multiple emails in textarea

AngularJS already supports the built-in validation with text input with type email. Something simple likes the following:
<input name="input" ng-model="email.text" required="" type="email" /> <span class="error" ng-show="myForm.input.$error.email"> Not valid email!</span>
However, I used a text area and I wanted to enter some email addresses that's saparated by a comma (,). I had a short research and it looked like AngualarJS has not supported this functionality so far. Therefore, I needed to build a custom directive that I could add my own validation functions. My validation was done only on client side, so I used the $validators object.

Note that, there is the $asyncValidators object which handles asynchronous validation, such as making an $http request to the backend.

This is just my implementation on my project. In order to understand that, I supposed you already had experiences with Angular…

The HelloWorld example of JSF 2.2 with Myfaces

I just did by myself create a very simple app "HelloWorld" of JSF 2.2 with a concrete implementation Myfaces that we can use it later on for our further JSF trying out. I attached the source code link at the end part. Just follow these steps below:

1. Create a Maven project in Eclipse (Kepler) with a simple Java web application archetype "maven-archetype-webapp". Maven should be the best choice for managing the dependencies, so far. JSF is a web framework that is the reason why I chose the mentioned archetype for my example.

2. Import dependencies for JSF implementation - Myfaces (v2.2.10) into file pom.xml. The following code that is easy to find from http://mvnrepository.com/ with key words "myfaces".

<dependency> <groupId>org.apache.myfaces.core</groupId> <artifactId>myfaces-api</artifactId> <version>2.2.10</version> </dependency> <dependency> <groupId>org.apache.myfaces.core</groupId&g…

My must-have apps for daily work

There is no doubt that cool apps can help us be more productive and enjoyable at work. For the time being, I really love the following apps which are used by me almost every day.
1. A personal Kanban In fact, a personal kanban is the most useful app for me. Why does it matter? It is not just a to-do list, but it keeps me motivated every day because it helps me be able to know what my "big picture" is. I usually set up my plans together with a path to reach them. KanbanFlow is my preferred tool.
2. A terminal Needless to say, a terminal is a must-have app for every developer, especially the ones use macOS/Linux. Due to its importance, I love to decorate and enhance it to be super exciting with various tools such as iTerm, oh-my-zsh, and thefuck. ;)

3. A documentation "ecosystem" As a developer, I can not remember all things that I have experimented a day. Moreover, a document is really useful for sharing an idea with other people. I use the set of tools for helping…

Strategy Design Pattern

For example, I have an program with an Animal abstract class and two sub-classes Dog and Bird. I want to add a new behavior for the class Animal, this is "fly". Now, I face to two approaches to solve this issue:

1. Adding an abstract method "fly" into the class Animal. Then, I force the sub-classes should be implemented this method, something like:

public abstract class Animal{ //bla bla public abstract void fly(); } public class Bird extends Animal{ //bla bla public void fly(){ System.out.println("Fly high"); } } public class Dog extends Animal{ //bla bla public void fly(){ System.out.println("Cant fly"); } }
2. Creating an interfaces with method "fly" inside. The same issue to abstract class, I force the classes these implement this interface should have a method "fly" inside:

public interface Flyable{ public void fly(); } public class Bird implements Flyable{ //bla bla public void fly(){ System.out.println…

vhandit | Python & Machine Learning

Search This Blog