I need to develop an image labeling application, for this task I'm considering using Conditional Random Fields (CRF) over a set of superpixels, there exists quite a few papers that point out this technology as the state of the art for this task. As usual the task could be devided into two tasks:
Training model: which for this problem would be obtaining the parameter vector 'w', using for example
Testing: which would be obtaining the most feasible label assignment of a given set of superpixels, i.e argmax(P(y|x))
I'm aware of training-time to be quite high, however I have not found anything about testing-time nor performance, does anyone have and idea of how much time could take the testing problem? I suppose it will depend on the number of labels, image size, implementation, hardware, etc
Testing is slowish because you still have to solve a graph cuts problem (but nothing like training). There is an implementation you can try out at http://drwn.anu.edu.au/drwnProjMultiSeg.html (you have probably seen Stephen Gould's papers).
I still have the log file. but it is a bit hard to interpret so the following may not be totally accurate. On a super fast machine, I think it took about:
4.5 hours cpu time to train 20 classes on 276 images from MSRC dataset
50 mins cpu time to classify 256 images, most of which was spent doing alpha expansion
Related
I am trying to perform DBSCAN on 18 million data points, so far just 2D but hoping to go up to 6D. I have not been able to find a way to run DBSCAN on that many points. The closest I got was 1 million with ELKI and that took an hour. I have used Spark before but unfortunately it does not have DBSCAN available.
Therefore, my first question is if anyone can recommend a way of running DBSCAN on this much data, likely in a distributed way?
Next, the nature of my data is that the ~85% lies in one huge cluster (anomaly detection). The only technique I have been able to come up with to allow me to process more data is to replace a big chunk of that huge cluster with one data point in a way that it can still reach all its neighbours (the deleted chunk is smaller than epsilon).
Can anyone provide any tips whether I'm doing this right or if there is a better way to reduce the complexity of DBSCAN when you know that most data is in one cluster centered around (0.0,0.0)?
Have you added an index to ELKI, and tried the parallel version? Except for the git version, ELKI will not automatically add an index; and even then fine-turning the index for the problem can help.
DBSCAN is not a good approach for anomaly detection - noise is not the same as anomalies. I'd rather use a density-based anomaly detection. There are variants that try to skip over "clear inliers" more efficiently if you know you are only interested in the top 10%.
If you already know that most of your data is in one huge cluster, why don't you directly model that big cluster, and remove it / replace it with a smaller approximation.
Subsample. There usually is next to no benefit to using the entire data. Even (or in particular) if you are interested in the "noise" objects, there is the trivial strategy of randomly splitting your data in, e.g., 32 subsets, then cluster each of these subsets, and join the results back. These 32 parts can be trivially processed in parallel on separate cores or computers; but because the underlying problem is quadratic in nature, the speedup will be anywhere between 32 and 32*32=1024.
This in particular holds for DBSCAN: larger data usually means you also want to use much larger minPts. But then the results will not differ much from a subsample with smaller minPts.
But by any means: before scaling to larger data, make sure your approach solves your problem, and is the smartest way of solving this problem. Clustering for anomaly detection is like trying to smash a screw into the wall with a hammer. It works, but maybe using a nail instead of a screw is the better approach.
Even if you have "big" data, and are proud of doing "big data", always begin with a subsample. Unless you can show that the result quality increases with data set size, don't bother scaling to big data, the overhead is too high unless you can prove value.
tl;dr - I use an autoencoder to try to reduce input dimensions for a reinforcement-learning (RL) agent to learn how to play Atari-KungFu. But it fails at encoding/decoding thrown knives, because they are only a couple pixels and getting them wrong probably has negligible impact on the autoencoder MSE loss (see green arrows in bottom left of image). This will probably permanently hobble the results. I want to figure out if there is a way to solve this -- preferably with a generalized solution, but I'd be happy for now with something specific to this problem.
Background:
I am working on Week5 of the "Practical Reinforcement Learning" course on Coursera (National Research University HSE), and I decided to spend extra time trying to expand performance on the Atari-KungFu assignment using Actor-Critic architecture. This post is not about actor-critic, but more about an interesting sub-problem I ran into related to autoencoders.
I create an encoder which outputs a tanh-64-neuron layer, which is used as a common input to the decoder, policy learner (actor), and value learner (critic). During training, the simulator returns batches of four sequential frames (64 x 144 x 4) and rewards from the last action. Then images are first used to train the autoencoder, then used with the rewards to train the actor & critic branches.
I display some metrics and example frames every 25000 iterations to see how it's doing. If the reconstructed images are accurate, then the inputs to the actor & critic branches should be getting good distilled information for efficient learning.
You can see below that the autoencoder is pretty good except for the thrown knives (see bottom-left). Arguably this is because missing those couple pixels minimally increases the MSE loss of the reconstructed image, so it has little incentive to learn it (and also there's not a lot of frames that have knives). Yet, seeing those knives is critical for the RL agent to learn to how to survive.
I haven't seen this kind of problem addressed before. A tiny artifact in the input images is crucial for learning, but is unlikely to be learned by the autoencoder. Can we fix/improve this?
IMO your problem is loss specific, some things which would probably help autoencoder reconstruct knife as well:
Find knives in input image using image processing techniques. Regions where knives are present should have higher loss value in MSE, say 10 times more. One way to find those semi-automatically could probably be convolution with big kernel; White pixels at the strict center would give more weight and only zeros around it would give it more weight as well. Something along these lines should find a region where only knives are located (throwing guys wouldn't, as they contain too many white pixels and holes). Using some threshold found empirically for the value of this kernel should be enough to correctly find them.
Lower loss for images when no knive was found, say divided by half. This would focus autoencoder harder on rarely seen cases when knive is seen.
On the downside - I suppose it could introduce some artifacts. In such case you may think about usage of pretrained encoder (like some version of ResNet) and increase model's capabilities.
I am working on a problem for which we aim to solve with deep Q learning. However, the problem is that training just takes too long for each episode, roughly 83 hours. We are envisioning to solve the problem within, say, 100 episode.
So we are gradually learning a matrix (100 * 10), and within each episode, we need to perform 100*10 iterations of certain operations. Basically we select a candidate from a pool of 1000 candidates, put this candidate in the matrix, and compute a reward function by feeding the whole matrix as the input:
The central hurdle is that the reward function computation at each step is costly, roughly 2 minutes, and each time we update one entry in the matrix.
All the elements in the matrix depend on each other in the long term, so the whole procedure seems not suitable for some "distributed" system, if I understood correctly.
Could anyone shed some lights on how we look at the potential optimization opportunities here? Like some extra engineering efforts or so? Any suggestion and comments would be appreciated very much. Thanks.
======================= update of some definitions =================
0. initial stage:
a 100 * 10 matrix, with every element as empty
1. action space:
each step I will select one element from a candidate pool of 1000 elements. Then insert the element into the matrix one by one.
2. environment:
each step I will have an updated matrix to learn.
An oracle function F returns a quantitative value range from 5000 ~ 30000, the higher the better (roughly one computation of F takes 120 seconds).
This function F takes the matrix as the input and perform a very costly computation, and it returns a quantitative value to indicate the quality of the synthesized matrix so far.
This function is essentially used to measure some performance of system, so it do takes a while to compute a reward value at each step.
3. episode:
By saying "we are envisioning to solve it within 100 episodes", that's just an empirical estimation. But it shouldn't be less than 100 episode, at least.
4. constraints
Ideally, like I mentioned, "All the elements in the matrix depend on each other in the long term", and that's why the reward function F computes the reward by taking the whole matrix as the input rather than the latest selected element.
Indeed by appending more and more elements in the matrix, the reward could increase, or it could decrease as well.
5. goal
The synthesized matrix should let the oracle function F returns a value greater than 25000. Whenever it reaches this goal, I will terminate the learning step.
Honestly, there is no effective way to know how to optimize this system without knowing specifics such as which computations are in the reward function or which programming design decisions you have made that we can help with.
You are probably right that the episodes are not suitable for distributed calculation, meaning we cannot parallelize this, as they depend on previous search steps. However, it might be possible to throw more computing power at the reward function evaluation, reducing the total time required to run.
I would encourage you to share more details on the problem, for example by profiling the code to see which component takes up most time, by sharing a code excerpt or, as the standard for doing science gets higher, sharing a reproduceable code base.
Not a solution to your question, just some general thoughts that maybe are relevant:
One of the biggest obstacles to apply Reinforcement Learning in "real world" problems is the astoundingly large amount of data/experience required to achieve acceptable results. For example, OpenAI in Dota 2 game colletected the experience equivalent to 900 years per day. In the original Deep Q-network paper, in order to achieve a performance close to a typicial human, it was required hundres of millions of game frames, depending on the specific game. In other benchmarks where the input are not raw pixels, such as MuJoCo, the situation isn't a lot better. So, if you don't have a simulator that can generate samples (state, action, next state, reward) cheaply, maybe RL is not a good choice. On the other hand, if you have a ground-truth model, maybe other approaches can easily outperform RL, such as Monte Carlo Tree Search (e.g., Deep Learning for Real-Time Atari Game Play Using Offline Monte-Carlo Tree Search Planning or Simple random search provides a competitive approach to reinforcement learning). All these ideas a much more are discussed in this great blog post.
The previous point is specially true for deep RL. The fact of approximatting value functions or policies using a deep neural network with millions of parameters usually implies that you'll need a huge quantity of data, or experience.
And regarding to your specific question:
In the comments, I've asked a few questions about the specific features of your problem. I was trying to figure out if you really need RL to solve the problem, since it's not the easiest technique to apply. On the other hand, if you really need RL, it's not clear if you should use a deep neural network as approximator or you can use a shallow model (e.g., random trees). However, these questions an other potential optimizations require more domain knowledge. Here, it seems you are not able to share the domain of the problem, which could be due a numerous reasons and I perfectly understand.
You have estimated the number of required episodes to solve the problem based on some empirical studies using a smaller version of size 20*10 matrix. Just a caution note: due to the curse of the dimensionality, the complexity of the problem (or the experience needed) could grow exponentially when the state space dimensionalty grows, although maybe it is not your case.
That said, I'm looking forward to see an answer that really helps you to solve your problem.
I am building an automatic translator in moses. To improve its performance, I use log-linear weight optimisation. This technique has a random component, which can affect slightly the final result (but I do not know exactly how much).
Suppose that the current performance of the model is 25 BLEU.
Suppose now I modify the language model (e.g. change the smoothing), and I get a performance of 26 BLEU.
My question is: how can I know if the improvement is because the modification, or is just noise from the random component?
This is pretty much what statistics is all about. You can basically do one of the two things (from the basic set of solutions, of course there are many more advanced):
try to measure/model/quantify the effect of randomness, if you know what is causing it, you might be able to actually compute how much it can affect your model. If analytical solution is not possible, you can always train 20 models with the same data/settings, gather results and estimate noise distribution. Once you have this you can perform statistical tests to check whether the improvement is statistically significant (for example by ANOVA tests).
simpler approach (but more expensive in terms of data/time) is to simply reduce the variance by averaging. In short - instead of training one model (or evaluating model once) which has this hard to determine noise component - do it many times, 10, 20, and average the results. This way you reduce the variance of the results in your analysis. This can (and should) be combined with the previous option - since now you have 20 results per run, thus you can again use statistical testes to see whether these are significantly different things.
Im wondering if there is a general rule of thumb for population sizing. Ive read in a book that 2x the chromosome length is a good starting point. Am i correct in assuming then that if i had an equation with 5 variables, i should have a population of 10?
Im also wondering if the following is correct:
Larger Population Size.
Pros:
Larger diversity so more likely to pick up on traits which return a good fitness.
Cons:
Requires longer to process.
vs
Smaller Population Size.
Pros:
Larger number of generations experienced per unit time.
Cons:
Mutation will have to be more prominent in order to compensate for smaller population??
EDIT
A little additional info, say i have an equation which has 5 unknown parameters. For each parameter i have anywhere between 10-50 values i would like to try assign to each of these variables. So for example
variable1 = 20 different values
variable2 = 15 different values
...
I thought a GA would be a decent approach to such a problem as the search space is quite large, ie worst case for the above would be 312,500,000 permutations (unless i have screwed up?) n!/(n-k)! where n = 50 and k = 1 => 50 * 50 * 50 * 50 * 50
unfortunately the number of parameters/range of values to check can vary alot so i was looking for some sort of rule of thumb as to how large i should set the population.
Thanks for ur help + if there is any more info you need/prefer to discuss in one of the chatrooms, just give me a shout.
I'm not sure where you read that 2x the chromosome length is a good starting point, but I'm guessing it's a book that concentrated on larger problems.
If you only have five variables, a genetic algorithm is probably not the right choice for converging upon a solution. With a chromosome length of five you're probably going to find that you very quickly reach a non-deterministic(this will change in subsequent runs) local minimum and slowly iterate around that space until you find the true local minimum.
However, if you are insistent on using a GA I would suggest abandoning that rule of thumb for this problem and really think about starting population as a measure of how far from the final solution you expect a random solution to be.
The reason that many rule of thumbs is dependent on chromosome length is because that's a decent proxy for this, if I have a hundred variables, and given randomly generating dna sequence is going to be further from ideal than if I had only one variable.
Additionally, if you're worried about computation intensity I'm going to go ahead and say that it shouldn't be an issue since you're dealing with such a small solution set. I think a better rule of thumb for smaller sets like this would be along the lines of:
(ln(chromosome_length*(solution_space/granularity)/mutation_rate))^2
Probably with a constant thrown in to scale for the particular problem.
It's definitely not a great rule of thumb (no rule is) but here's my logic for it:
Chromosome length is just a proxy for size of solution space, so taking into account the size of the solution space will necessarily increase the accuracy of this proxy
A smaller mutation rate necessitates a larger population size to compensate for the fact that you are more prone to get caught in local minima
Any rule of thumb should scale logarithmically since a genetic algorithm is akin to a tree search of your solution space.
The squared term was mostly the result of trying this out, but it looks like the logarithmic scaling was a little aggressive, though the general shape seemed right.
However I think a better choice would be to start at a reasonable number (100) and try iterating up and down until you find a population size that seems to balance accuracy with execution speed.
As with most genetic algorithm parameters population size is highly dependant on the problem. There are certain factors that can help to point in the direction of whether you should have a large or small population size but a lot of the time testing different values against a known solution before running it on your problem is a good idea (if this is possible of course).
A population size of 10 does seem rather small though. You say you have an equation with five variables. Is your problem represented by a chromosome of 5 values? It seems small for a chromosome and if this is the case it's likely that using a genetic algorithm may not be the best way to solve the problem. Perhaps if you give a bit more detail on your problem and how you are representing it people may have a better idea of how to advise you.
I'd also add that your cons for large and small population sizes aren't exactly correct. A larger population size does take longer to process than a small one but since it can often solve the problem quicker then overall the processing time isn't necessarily longer. gain, it's highly dependant on the problem. With a smaller population size mutation shouldn't have to be more prominent. Mutation is generally used to stop the genetic algorithm from becoming stuck in a local maximum and should usually be a very small value. A small population is more likely to become stuck in a local maximum but if you have a mutation value which is too high you may be nullifying the natural improvement of the genetic algorithm.