Can anybody please explain multitask learning in simple and intuitive way? May be some real
world problem would be useful.Mostly, these days i am seeing many people are using it for natural language processing tasks.
Let's say you've built a sentiment classifier for a few different domains. Say, movies, music DVDs, and electronics. These are easy to build high quality classifiers for, because there is tons of training data that you've scraped from Amazon. Along with each classifier, you also build a similarity detector that will tell you for a given piece of text, how similar it was to the dataset each of the classifiers was trained on.
Now you want to find the sentiment of some text from an unknown domain or one in which there isn't such a great dataset to train on. Well, how about we take a similarity weighted combination of the classifications from the three high quality classifiers we already have. If we are trying to classify a dish washer review (there is no giant corpus of dish washer reviews, unfortunately), it's probably most similar to electronics, and so the electronics classifier will be given the most weight. On the other hand, if we are trying to classify a review of a TV show, probably the movies classifier will do the best job.
Related
Currently, I'm working on a project where I need to extract the relevant aspects used in positive and negative reviews in real time.
For the notions of more negative and positive, it will be a question of contextualizing the word. Distinguish between a word that sounds positive in a negative context (consider irony).
Here is an example:
Very nice welcome!!! We ate very well with traditional dishes as at home, the quality but also the quantity are in appointment!!!*
Positive aspects: welcome, traditional dishes, quality, quantity
Can anyone suggest to me some tutorials, papers or ideas about this topic?
Thank you in advance.
This task is called Aspect Based Sentiment Analysis (ABSA). Most popular is the format and dataset specified in the 2014 Semantic Evaluation Workshop (Task 5) and its updated versions in the following years.
Overview of model efficiencies over the years:
https://paperswithcode.com/sota/aspect-based-sentiment-analysis-on-semeval
Good source for ressources and repositories on the topic (some are very advanced but there are some more starter friendly ressources in there too):
https://github.com/ZhengZixiang/ABSAPapers
Just from my general experience in this topic a very powerful starting point that doesn't require advanced knowledge in machine learning model design is to prepare a Dataset (such as the one provided for the SemEval2014 Task) that is in a Token Classification Format and use it to fine-tune a pretrained transformer model such as BERT, RoBERTa or similar. Check out any tutorial on how to do fine-tuning on a token classification model like this one in huggingface. They usually use the popular task of Named Entity Recognition (NER) as the example task but for the ABSA-Task you basically do the same thing but with other labels and a different dataset.
Obviously an even easier approach would be to take more rule-based approaches or combine a rule-based approach with a trained sentiment analysis model/negation detection etc., but I think generally with a rule-based approach you can expect a much inferior performance compared to using state-of-the-art models as transformers.
If you want to go even more advanced than just fine-tuning the pretrained transformer models then check out the second and third link I provided and look at some of the machine learning model designs specifically designed for Aspect Based Sentiment Analysis.
I am given a task of classifying a given news text data into one of the following 5 categories - Business, Sports, Entertainment, Tech and Politics
About the data I am using:
Consists of text data labeled as one of the 5 types of news statement (Bcc news data)
I am currently using NLP with nltk module to calculate the frequency distribution of every word in the training data with respect to each category(except the stopwords).
Then I classify the new data by calculating the sum of weights of all the words with respect to each of those 5 categories. The class with the most weight is returned as the output.
Heres the actual code.
This algorithm does predict new data accurately but I am interested to know about some other simple algorithms that I can implement to achieve better results. I have used Naive Bayes algorithm to classify data into two classes (spam or not spam etc) and would like to know how to implement it for multiclass classification if it is a feasible solution.
Thank you.
In classification, and especially in text classification, choosing the right machine learning algorithm often comes after selecting the right features. Features are domain dependent, require knowledge about the data, but good quality leads to better systems quicker than tuning or selecting algorithms and parameters.
In your case you can either go to word embeddings as already said, but you can also design your own custom features that you think will help in discriminating classes (whatever the number of classes is). For instance, how do you think a spam e-mail is often presented ? A lot of mistakes, syntaxic inversion, bad traduction, punctuation, slang words... A lot of possibilities ! Try to think about your case with sport, business, news etc.
You should try some new ways of creating/combining features and then choose the best algorithm. Also, have a look at other weighting methods than term frequencies, like tf-idf.
Since your dealing with words I would propose word embedding, that gives more insights into relationship/meaning of words W.R.T your dataset, thus much better classifications.
If you are looking for other implementations of classification you check my sample codes here , these models from scikit-learn can easily handle multiclasses, take a look here at documentation of scikit-learn.
If you want a framework around these classification that is easy to use you can check out my rasa-nlu, it uses spacy_sklearn model, sample implementation code is here. All you have to do is to prepare the dataset in a given format and just train the model.
if you want more intelligence then you can check out my keras implementation here, it uses CNN for text classification.
Hope this helps.
I'm trying to implement as aspect miner based on consumer reviews in amazon for durable- washing machine, refrigerator. The idea is to output sentiment polarity for aspects instead of the entire sentence. For eg: 'Food was good but service was bad' review must output food to be positive and service to be negative. I read through Richard Socher's paper on RNTN model for fine grained sentiment classifier but I guess I'll need to manually tag sentiment for phrases for a different domain and create my own treebank for better accuracy.
Here's an alternate approach I'd thought of. Could someone pls validate/guide me with your feedback
Break the approach into 2 sub tasks. 1) Identify aspects 2) Identify sentiment
Identify aspects
Use POS tagger to identify all nouns. This should shortlist
potentially all aspects in the reviews.
Use word2vec of these nouns to determine similar nouns and reduce the dataset size
Identify sentiments
Train a CNN or dense net model on reviews with rating 1,2,4,5(ignore
3 as we need data that has polarity)
Breakdown the test set reviews into phrases(eg 'Food was good') and then score them using the above model
Find the aspects identified in the 1st sub task and tag them to
their respective phrases.
I don't know how to answer this question but have a few suggestions:
Take a look at multitask learning in neuralnets literature and try an end2end neuralnet for multiple tasks.
Use pretrained word vectors like w2v or glov as inputs.
Don't rely on pos taggers when you use internet data,
Find a way to represent your name entities and oov in your design.
Don't ignore 3!!
You should annotate some data periodically.
I just wanted to understand (from your experience), that if I have to create a sentiment analysis classification model (using NLTK), what would be a good training data size. For instance if my training data is going to contain tweets, and I intend to classify them as positive,negative and neutral, how many tweets each should I ideally have per category to get a reasonable model working?
I understand that there are many parameters like quality of data, but if one has to get started what might be a good number.
That's a really hard question to answer for people who are not familiar with the exact data, its labelling and the application you want to use it for. But as a ballpark estimate, I would say start with 1,000 examples of each and go from there.
I am about to start a project where my final goal is to classify short texts into classes: "may be interested in visiting place X" : "not interested or neutral". Place is described by set of keywords (e.g. meals or types of miles like "chinese food"). So ideally I need some approach to model desire of user based on short text analysis - and then classify based on a desire score or desire probability - is there any state-of-the-art in this field ? Thank you
This problem is exactly the same as sentiment analysis of texts. But, instead of the traditional binary classification, you seem to have a "neutral" opinion. State-of-the-art in sentiment analysis is highly domain-dependent. Techniques that have excelled in classifying movies do not perform as well on commercial products, for example.
Additionally, even the feature-selection is highly domain-dependent. For example, unigrams work well for movie review classification, but a combination of unigrams and bigrams perform better for classifying twitter texts.
My best advice is to "play around" with different features. Since you are looking at short texts, twitter is probably a good motivational example. I would start with unigrams and bigrams as my features. The exact algorithm is not very important. SVM usually performs very well with correct parameter tuning. Use a small amount of held-out data for tuning these parameters before experimenting on bigger datasets.
The more interesting portion of this problem is the ranking! A "purity score" has been recently used for this purpose in the following papers (and I'd say they are pretty state-of-the-art):
Sentiment summarization: evaluating and learning user preferences. Lerman, Blair-Goldensohn and McDonald. EACL. 2009.
The viability of web-derived polarity lexicons. Velikovich, Blair-Goldensohn, Hannan and McDonald. NAACL. 2010.