I am using word2vec (and doc2vec) to get embeddings for sentences, but i want to completely ignore word order.
I am currently using gensim, but can use other packages if necessary.
As an example, my text looks like this:
[
['apple', 'banana','carrot','dates', 'elderberry', ..., 'zucchini'],
['aluminium', 'brass','copper', ..., 'zinc'],
...
]
I intentionally want 'apple' to be considered as close to 'zucchini' as it is to 'banana' so I have set the window size to a very large number, say 1000.
I am aware of 2 problems that may arise with this.
Problem 1:
The window might roll in at the start of a sentence creating the following training pairs:
('apple', ('banana')), ('apple', ('banana', 'carrot')), ('apple', ('banana', 'carrot', 'date')) before it eventually gets to the correct ('apple', ('banana','carrot', ..., 'zucchini')).
This would seem to have the effect of making 'apple' closer to 'banana' than 'zucchini',
since their are so many more pairs containing 'apple' and 'banana' than there are pairs containing 'apple' and 'zucchini'.
Problem 2:
I heard that pairs are sampled with inverse proportion to the distance from the target word to the context word- This also causes an issue making nearby words more seem more connected than I want them to be.
Is there a way around problems 1 and 2?
Should I be using cbow as opposed to sgns? Are there any other hyperparameters that I should be aware of?
What is the best way to go about removing/ignoring the order in this case?
Thank you
I'm not sure what you mean by "Problem 1" - there's no "roll" or "wraparound" in the usual interpretation of a word2vec-style algorithm's window parameter. So I wouldn't worry about this.
Regarding "Problem 2", this factor can be essentially made negligible by the choice of a giant window value – say for example, a value one million times larger than your largest sentence. Then, any difference in how the algorithm treats the nearest-word and the 2nd-nearest-word is vanishingly tiny.
(More specifically, the way the gensim implementation – which copies the original Google word2vec.c in this respect – achieves a sort of distance-based weighting is actually via random dynamic shrinking of the actual window used. That is, for each visit during training to each target word, the effective window truly used is some random number from 1 to the user-specified window. By effectively using smaller windows much of the time, the nearer words have more influence – just without the cost of performing other scaling on the whole window's words every time. But in your case, with a giant window value, it will be incredibly rare for the effective-window to ever be smaller than your actual sentences. Thus every word will be included, equally, almost every time.)
All these considerations would be the same using SG or CBOW mode.
I believe a million-times-larger window will be adequate for your needs, for if for some reason it wasn't, another way to essentially cancel-out any nearness effects could be to ensure your corpus's items individual word-orders are re-shuffled between each time they're accessed as training data. That ensures any nearness advantages will be mixed evenly across all words – especially if each sentence is trained on many times. (In a large-enough corpus, perhaps even just a 1-time shuffle of each sentence would be enough. Then, over all examples of co-occurring words, the word co-occurrences would be sampled in the right proportions even with small windows.)
Other tips:
If your training data starts in some arranged order that clumps words/topics together, it can be beneficial to shuffle them into a random order instead. (It's better if the full variety of the data is interleaved, rather than presented in runs of many similar examples.)
When your data isn't true natural-language data (with its usual distributions & ordering significance), it may be worth it to search further from the usual defaults to find optimal metaparameters. This goes for negative, sample, & especially ns_exponent. (One paper has suggested the optimal ns_exponent for training vectors for recommendation-systems is far different from the usual 0.75 default for natural-language modeling.)
Related
I am trying to assess a doc2vec model based on the code from here. Basically, I want to know the percentual of inferred documents are found to be most similar to itself. This is my current code an:
for doc_id, doc in enumerate(cur.execute('SELECT Text FROM Patents')):
docs += 1
doc = clean_text(doc)
inferred_vector = model.infer_vector(doc)
sims = model.docvecs.most_similar([inferred_vector], topn=len(model.docvecs))
rank = [docid for docid, sim in sims].index(doc_id)
ranks.append(rank)
counter = collections.Counter(ranks)
accuracy = counter[0] / docs
This code works perfectly with smaller datasets. However, since I have a huge file with millions of documents, this code becomes too slow, it would take months to compute. I profiled my code and most of the time is consumed by the following line: sims = model.docvecs.most_similar([inferred_vector], topn=len(model.docvecs)).
If I am not mistaken, this is having to measure each document to every other document. I think computation time might be massively reduced if I change this to topn=1 instead since the only thing I want to know is if the most similar document is itself or not. Doing this will basically take each doc (i.e., inferred_vector), measure its most similar document (i.e., topn=1), and then I just see if it is itself or not. How could I implement this? Any help or idea is welcome.
To have most_similar() return only the single most-similar document, that is as simple as specifying topn=1.
However, to know which one document of the millions is the most-similar to a single target vector, the similarities to all the candidates must be calculated & sorted. (If even one document was left out, it might have been the top-ranked one!)
Making sure absolutely no virtual-memory swapping is happening will help ensure that brute-force comparison happens as fast as possible, all in RAM – but with millions of docs, it will still be time-consuming.
What you're attempting is a fairly simple "self-check" as to whether training led to self-consistent model: whether the re-inference of a document creates a vector "very close to" the same doc-vector left over from bulk training. Failing that will indicate some big problems in doc-prep or training, but it's not a true measure of the model's "accuracy" for any real task, and the model's value is best evaluated against your intended use.
Also, because this "re-inference self-check" is just a crude sanity check, there's no real need to do it for every document. Picking a thousand (or ten thousand, or whatever) random documents will give you a representative idea of whether most of the re-inferred vectors have this quality, or not.
Similarly, you could simply check the similarity of the re-inferred vector against the single in-model vector for that same document-ID, and check whether they are "similar enough". (This will be much faster, but could also be done on just a random sample of docs.) There's no magic proper threshold for "similar enough"; you'd have to pick one that seems to match your other goals. For example, using scikit-learn's cosine_similarity() to compare the two vectors:
from sklearn.metrics.pairwise import cosine_similarity
# ...
inferred_vector = model.infer_vector(doc_words)
looked_up_vector = model.dv[doc_id]
self_similarity = cosine_similarity([inferred_vector], [looked_up_vector])[0]
# then check that value against some threshold
(You have to wrap the single vectors in lists as arguments to cosine_similarity(), then access the 0th element of the return value, because it is designed to usually work on larger lists of vectors.)
With this calculation, you wouldn't know if, for example, some of the other stored-doc-vectors are a little closer to your inferred target - but that may not be that important, anyway. The docs might be really similar! And while the original "closest to itself" self-check will fail miserably if there were major defects in training, even a well-trained model will likely have some cases where natural model jitter prevents a "closest to itself" for every document. (With more documents inside the same number of dimensions, or certain corpuses with lots of very-similar documents, this would become more common... but not be a concerning indicator of any model problems.)
I'm trying to build a Word2vec (or FastText) model using Gensim on a massive dataset which is composed of 1000 files, each contains ~210,000 sentences, and each sentence contains ~1000 words. The training was made on a 185gb RAM, 36-core machine.
I validated that
gensim.models.word2vec.FAST_VERSION == 1
First, I've tried the following:
files = gensim.models.word2vec.PathLineSentences('path/to/files')
model = gensim.models.word2vec.Word2Vec(files, workers=-1)
But after 13 hours I decided it is running for too long and stopped it.
Then I tried building the vocabulary based on a single file, and train based on all 1000 files as follows:
files = os.listdir['path/to/files']
model = gensim.models.word2vec.Word2Vec(min_count=1, workers=-1)
model.build_vocab(corpus_file=files[0])
for file in files:
model.train(corpus_file=file, total_words=model.corpus_total_words, epochs=1)
But I checked a sample of word vectors before and after the training, and there was no change, which means no actual training was done.
I can use some advise on how to run it quickly and successfully. Thanks!
Update #1:
Here is the code to check vector updates:
file = 'path/to/single/gziped/file'
total_words = 197264406 # number of words in 'file'
total_examples = 209718 # number of records in 'file'
model = gensim.models.word2vec.Word2Vec(iter=5, workers=12)
model.build_vocab(corpus_file=file)
wv_before = model.wv['9995']
model.train(corpus_file=file, total_words=total_words, total_examples=total_examples, epochs=5)
wv_after = model.wv['9995']
so the vectors: wv_before and wv_after are exactly the same
There's no facility in gensim's Word2Vec to accept a negative value for workers. (Where'd you get the idea that would be meaningful?)
So, it's quite possible that's breaking something, perhaps preventing any training from even being attempted.
Was there sensible logging output (at level INFO) suggesting that training was progressing in your trial runs, either against the PathLineSentences or your second attempt? Did utilities like top show busy threads? Did the output suggest a particular rate of progress & let you project-out a likely finishing time?
I'd suggest using a positive workers value and watching INFO-level logging to get a better idea what's happening.
Unfortunately, even with 36 cores, using a corpus iterable sequence (like PathLineSentences) puts gensim Word2Vec in a model were you'll likely get maximum throughput with a workers value in the 8-16 range, using far less than all your threads. But it will do the right thing, on a corpus of any size, even if it's being assembled by the iterable sequence on-the-fly.
Using the corpus_file mode can saturate far more cores, but you should still specify the actual number of worker threads to use – in your case, workers=36 – and it is designed to work on from a single file with all data.
Your code which attempts to train() many times with corpus_file has lots of problems, and I can't think of a way to adapt corpus_file mode to work on your many files. Some of the problems include:
you're only building the vocabulary from the 1st file, which means any words only appearing in other files will be unknown and ignored, and any of the word-frequency-driven parts of the Word2Vec algorithm may be working on unrepresentative
the model builds its estimate of the expected corpus size (eg: model.corpus_total_words) from the build_vocab() step, so every train() will behave as if that size is the total corpus size, in its progress-reporting & management of the internal alpha learning-rate decay. So those logs will be wrong, the alpha will be mismanaged in a fresh decay each train(), resulting in a nonsensical jigsaw up-and-down alpha over all files.
you're only iterating over each file's contents once, which isn't typical. (It might be reasonable in a giant 210-billion word corpus, though, if every file's text is equally and randomly representative of the domain. In that case, the full corpus once might be as good as iterating over a corpus that's 1/5th the size 5 times. But it'd be a problem if some words/patterns-of-usage are all clumped in certain files - the best training interleaves contrasting examples throughout each epoch and all epochs.)
min_count=1 is almost always unwise with this algorithm, and especially so in large corpora of typical natural-language word frequencies. Rare words, and especially those appearing only once or a few times, make the model gigantic but those words won't get good word-vectors, and keeping them in acts like noise interfering with the improvement of other more-common words.
I recommend:
Try the corpus iterable sequence mode, with logging and a sensible workers value, to at least get an accurate read of how long it might take. (The longest step will be the initial vocabulary scan, which is essentially single-threaded and must visit all data. But you can .save() the model after that step, to then later re-.load() it, tinker with settings, and try different train() approaches without repeating the slow vocabulary survey.)
Try aggressively-higher values of min_count (discarding more rare words for a smaller model & faster training). Perhaps also try aggressively-smaller values of sample (like 1e-05, 1e-06, etc) to discard a larger fraction of the most-frequent words, for faster training that also often improves overall word-vector quality (by spending relatively more effort on less-frequent words).
If it's still too slow, consider if you could using a smaller subsample of your corpus might be enough.
Consider the corpus_file method if you can roll much or all of your data into the single file it requires.
I'm currently having a problem with the conception of an algorithm.
I want to create a WYSIWYG editor that goes along the current [bbcode] editor I have.
To do that, I use a div with contenteditable set to true for the WYSIWYG editor and a textarea containing the associated bbcode. Until there, no problem. But my concern is that if a user wants to add a tag (for example, the [b] tag), I need to know where they want to include it.
For that, I need to know exactly where in the bbcode I should insert the tags. I thought of comparing the two texts (one with html tags like <span>, the other with bbcode tags like [b]), and that's where I'm struggling.
I did some research but couldn't find anything that would help me, or I did not understand it correctly (maybe did I do a wrong research). What I could find is the Jaccard index, but I don't really know how to make it work correctly.
I also thought of another alternative. I could just take the code in the WYSIWYG editor before the cursor location, and split it every time I encounter a html tag. That way, I can, in the bbcode editor, search for the first occurrence, then search for the second occurrence starting at the last index found, and so on until I reach the place where the cursor is pointing at.
I'm not sure if it would work, and I find that solution a bit dirty. Am I totally wrong or should I do it this way?
Thanks for the help.
A popular way of determining what is the level of the similarity between the two texts is computing the mentioned Jaccard similarity. Citing Wikipedia:
The Jaccard index, also known as Intersection over Union and the Jaccard similarity coefficient, is a statistic used for comparing the similarity and diversity of sample sets. The Jaccard coefficient measures the similarity between finite sample sets, and is defined as the size of the intersection divided by the size of the union of the sample sets:
If you have a large number of texts though, computing the full Jaccard index of every possible combination of two texts is super computationally expensive. There is another way to approximate this index that is called minhashing. What it does is use several (e.g. 100) independent hash functions to create a signature and it repeats this procedure many times. This whole process has a nice property that the probability (over all permutations) that T1 = T2 is the same as J(A,B).
Another way to cluster similar texts (or any other data) together is to use Locality Sensitive Hashing which by itself is an approximation of what KNN does, and is usually worse than that, but is definitely faster to compute. The basic idea is to project the data into low-dimensional binary space (that is, each data point is mapped to a N-bit vector, the hash key). Each hash function h must satisfy the sensitive hashing property prob[h(x)=h(y)]=sim(x,y) where sim(x,y) in [0,1] is the similarity function of interest. For dots products it can be visualized as follows:
we can now ask what would be the has of the indicated point (in this case it's 101) and everything that is close to this point has the same hash.
EDIT to answer the comment
No, you asked about the text similarity and so I answered that. You basically ask how can you predict the position of the character in text 2. It depends on whether you analyze the writer's style or just pure syntax. In any of those two cases, IMHO you need some sort of statistics that will tell where it is likely for this character to occur given all the other data/text. You can go with n-grams, RNNs, LSTMs, Markov Chains or any other form of sequential data analysis.
I am training multiple word2vec models with Gensim. Each of the word2vec will have the same parameter and dimension, but trained with slightly different data. Then I want to compare how the change in data affected the vector representation of some words.
But every time I train a model, the vector representation of the same word is wildly different. Their similarity among other words remain similar, but the whole vector space seems to be rotated.
Is there any way I can rotate both of the word2vec representation in such way that same words occupy same position in vector space, or at least they are as close as possible.
Thanks in advance.
That the locations of words vary between runs is to be expected. There's no one 'right' place for words, just mutual arrangements that are good at the training task (predicting words from other nearby words) – and the algorithm involves random initialization, random choices during training, and (usually) multithreaded operation which can change the effective ordering of training examples, and thus final results, even if you were to try to eliminate the randomness by reliance on a deterministically-seeded pseudorandom number generator.
There's a class called TranslationMatrix in gensim that implements the learn-a-projection-between-two-spaces method, as used for machine-translation between natural languages in one of the early word2vec papers. It requires you to have some words that you specify should have equivalent vectors – an anchor/reference set – then lets other words find their positions in relation to those. There's a demo of its use in gensim's documentation notebooks:
https://github.com/RaRe-Technologies/gensim/blob/develop/docs/notebooks/translation_matrix.ipynb
But, there are some other techniques you could also consider:
transform & concatenate the training corpuses instead, to both retain some words that are the same across all corpuses (such as very frequent words), but make other words of interest different per segment. For example, you might leave words like "hot" and "cold" unchanged, but replace words like "tamale" or "skiing" with subcorpus-specific versions, like "tamale(A)", "tamale(B)", "skiing(A)", "skiing(B)". Shuffle all data together for training in a single session, then check the distances/directions between "tamale(A)" and "tamale(B)" - since they were each only trained by their respective subsets of the data. (It's still important to have many 'anchor' words, shared between different sets, to force a correlation on those words, and thus a shared influence/meaning for the varying-words.)
create a model for all the data, with a single vector per word. Save that model aside. Then, re-load it, and try re-training it with just subsets of the whole data. Check how much words move, when trained on just the segments. (It might again help comparability to hold certain prominent anchor words constant. There's an experimental property in the model.trainables, with a name ending _lockf, that lets you scale the updates to each word. If you set its values to 0.0, instead of the default 1.0, for certain word slots, those words can't be further updated. So after re-loading the model, you could 'freeze' your reference words, by setting their _lockf values to 0.0, so that only other words get updated by the secondary training, and they're still bound to have coordinates that make sense with regard to the unmoving anchor words. Read the source code to better understand how _lockf works.)
I've seen some questions on deciding the best OCR result given output from different engines, and the answer is typically "choose the best engine".
I want, however, to capture several frames of text images, with possible temporary occlusions or temporary failures.
I'm using tesseract-ocr with python-tesseract.
Considering the OCR outputs of the last N frames, I want to decide what is the best result (line by line, for simplicity).
For example, for N=3, we could use a median filtering:
ABXD
XBCX
AXCD
When there are 2 out of 3 equal characters, the majority will win, so the result would be ABCD.
However, that's not so easy with different string sizes. If I expect a given size M (if scanning a price table, the rows are typically XX.XX), I can always penalize on strings bigger than M.
If we were talking numbers, a median filtering would work quite well (simple background subtraction in computer vision), or some least mean squares adaptive filtering.
There's also the problem of similar characters: l and 1 can be very similar, depending on the font.
I was also thinking of using string distances between each string. For example, choose the string with the smallest sum of distances with the others.
Has anyone addressed this kind of problem before? Is there any known algorithm for this kind of problem that I should know?
This problem is called multiple sequence alignment and you can read about it here