Unsupervised automatic tagging algorithms?

Question

I want to build a web application that lets users upload documents, videos, images, music, and then give them an ability to search them. Think of it as Dropbox + Semantic Search.

When user uploads a new file, e.g. Document1.docx, how could I automatically generate tags based on the content of the file? In other words no user input is needed to determine what the file is about. If suppose that Document1.docx is a research paper on data mining, then when user searches for data mining, or research paper, or document1, that file should be returned in search results, since data mining and research paper will most likely be potential auto-generated tags for that given document.

1. Which algorithms would you recommend for this problem?

2. Is there an natural language library that could do this for me?

3. Which machine learning techniques should I look into to improve tagging precision?

4. How could I extend this to video and image automatic tagging?

Thanks in advance!

Marcellmarcella

How would you search for a video? Would you supply another video or would you (more naturally) input a few words describing it. If the latter you're going to need some sort of user involvement in tagging. – Polychrome

I am pretty sure you can get a lot of literature by googling it. Because as far as I know, there are even a bunch of research works out there about trying to tag videos automatically. Given that text is much easier for machines to interpret than videos or images, I believe you can find what you want on the website. But keep in mind, there is no perfect algorithm that can do the things exactly what you expect. – Wichern

@jozefg Two options in my mind right now: a) Either input a few keywords b) Extract audio channel, analyze it for patterns. If speech, parse speech to text and extract relevant keywords. If music, pass it to Echospirit for music identification. All other cases will result in no tags. – Marcellmarcella

In other words, you want to build Google. I commend ambitious projects. – Stricture

Were you able to make it? – Naquin

Answer 1

The most common unsupervised machine learning model for this type of task is Latent Dirichlet Allocation (LDA). This model automatically infers a collection of topics over a corpus of documents based on the words in those documents. Running LDA on your set of documents would assign words with probability to certain topics when you search for them, and then you could retrieve the documents with the highest probabilities to be relevant to that word.

There have been some extensions to images and music as well, see http://cseweb.ucsd.edu/~dhu/docs/research_exam09.pdf.

LDA has several efficient implementations in several languages:

• many implementations from the original researchers
• http://mallet.cs.umass.edu/, written in Java and recommended by others on SO
• PLDA: a fast, parallelized C++ implementation

Cupola

As much as I am an LDA supporter, i dont think it the "topics" generated from LDA has any value to produce any useable tags other than for WSI purpose. After generating the topics, an intermediate step to map topics to a set of semantic/syntactic annotation is required to make a knowledge-driven NLP annotation, especially one that previous researches had put so much consideration to create. – Jointed

I agree with you, but I think that searching for words that are associated with certain topics can also retrieve documents with high priors on certain topics that are similar. The OP will have to go somewhere else for a much more rigorous treatment of this :) – Cupola

Answer 2

These guys propose an alternative to LDA.

Automatic Tag Recommendation Algorithms for Social Recommender Systems http://research.microsoft.com/pubs/79896/tagging.pdf

Haven't read thru the whole paper but they have two algorithms:

1. Supervised learning version. This isn't that bad. You can use Wikipedia to train the algorithm
2. "Prototype" version. Haven't had a chance to go thru this but this is what they recommend

UPDATE: I've researched this some more and I've found another approach. Basically, it's a two-stage approach that's very simple to understand and implement. While too slow for 100,000s of documents, it (probably) has good performance for 1000s of docs (so it's perfect for tagging a single user's documents). I'm going to try this approach and will report back on performance/usability.

In the mean time, here's the approach:

1. Use TextRank as per http://qr.ae/36RAP to generate a tag list for a single document. This generates a tag list for a single document independent of other documents.
2. Use the algorithm from "Using Machine Learning to Support Continuous Ontology Development" (https://www.researchgate.net/publication/221630712_Using_Machine_Learning_to_Support_Continuous_Ontology_Development) to integrate the tag list (from step 1) into the existing tag list.

Ferricyanide

No update then? The answers here are probably outdated by now written 4 years ago – Schechter

Answer 3

Text documents can be tagged using this keyphrase extraction algorithm/package. http://www.nzdl.org/Kea/ Currently it supports limited type of documents (Agricultural and medical I guess) but you can train it according to your requirements.

I'm not sure how would the image/video part work out, unless you're doing very accurate object detection (which has it's own shortcomings). How are you planning to do it ?

Militarize

If you have a question/need more information from the OP, you should first post a comment to the question to get more information, then post an answer that you know will be relevant. – Smythe

Answer 4

You want Doc-Tags (https://www.Doc-Tags.com) which is a commercial product that automatically and Unsupervised - generates Contextually Accurate Document Tags. The built-in Reporting functionality makes the product a light-weight document management system.

For Developers wanting to customize their own approach - the source code is available (very cheap) and the back-end service xAIgent (https://xAIgent.com) is very inexpensive to use.

Stele

Answer 5

I posted a blog article today to answer your question.

http://scottge.net/2015/06/30/automatic-image-and-video-tagging/

There are basically two approaches to automatically extract keywords from images and videos.

1. Multiple Instance Learning (MIL)
2. Deep Neural Networks (DNN), Recurrent Neural Networks (RNN), and the variants

In the above blog article, I list the latest research papers to illustrate the solutions. Some of them even include demo site and source code.

Thanks, Scott

Diagonal