1. Get Another Label? Improving Data Quality and Machine Learning Using Multiple, Noisy Labelers Panos Ipeirotis New York University Joint work with Jing Wang, Foster Provost, and Victor Sheng

3. 3

4. 4

5. 5

6. 6 Outsourcing machine learning preprocessing Traditionally, modeling teams have invested substantial internal resources in data formulation, information extraction, cleaning, and other preprocessing Now, we can outsource preprocessing tasks, such as labeling, feature extraction, verifying information extraction, etc. – using Mechanical Turk, oDesk, etc.Mechanical Turk – quality may be lower than expert labeling (much?) – but low costs can allow massive scale

7. Example: Build an “Adult Web Site” Classifier Need a large number of hand-labeled sites Get people to look at sites and classify them as: G (general audience) PG (parental guidance) R (restricted) X (porn)

8. 8

9. Example: Build an “Adult Web Site” Classifier Need a large number of hand-labeled sites Get people to look at sites and classify them as: G (general audience) PG (parental guidance) R (restricted) X (porn) Cost/Speed Statistics  Undergrad intern: 200 websites/hr, cost: $15/hr  MTurk: 2500 websites/hr, cost: $12/hr Cost/Speed Statistics  Undergrad intern: 200 websites/hr, cost: $15/hr  MTurk: 2500 websites/hr, cost: $12/hr

10. 10 Noisy labels can be problematic Many tasks rely on high-quality labels for objects: – webpage classification for safe advertising – learning predictive models – searching for relevant information – finding duplicate database records – image recognition/labeling – song categorization – sentiment analysis Noisy labels can lead to degraded task performance

11. 11 Quality and Classification Performance Labeling quality increases  classification quality increases P = 50% P = 60% P = 80% P = 100% P: Single-labeler quality (probability of assigning correctly a binary label)

12. Solutions Get better labelers – Often beyond our control or too expensive Get more labelers per item – Our focus 12

13. 13 Majority Voting and Label Quality P=0.4 P=0.5 P=0.6 P=0.7 P=0.8 P=0.9 P=1.0 Ask multiple labelers, keep majority label as “true” label Quality is probability of being correct P is probability of individual labeler being correct

14. 14 Tradeoffs for Modeling Get more examples  Improve classification Get more labels  Improve label quality  Improve classification P = 0.5 P = 0.6 P = 0.8 P = 1.0

15. 15 Basic Labeling Strategies Single Labeling – Get as many data points as possible – One label each Round-robin Repeated Labeling – Repeatedly label data points, giving next label to the one with the fewest so far

16. 16 Round Robin vs. Single: Low Noise p= 0.8, labeling quality #examples =50 FRR (50 examples) Single With low noise/few examples, more (single labeled) examples better

17. 17 Round Robin vs. Single: High Noise p= 0.6, labeling quality #examples =100 FRR (100 examples) SL High noise/many examples, repeated labeling better

18. 18 Selective Repeated-Labeling We have seen so far: – With enough examples and noisy labels, getting multiple labels is better than single-labeling – When we consider extra cost for getting the unlabeled part the benefit is magnified Can we do better than the basic strategies? Key observation: we have additional information to guide selection of data for repeated labeling  the current multiset of labels Example: {+,-,+,-,-,+} vs. {+,+,+,+,+,+}

19. 19 Natural Candidate: Entropy Entropy is a natural measure of label uncertainty: E({+,+,+,+,+,+})=0 E({+,-, +,-, -,+ })=1 Strategy: Get more labels for high-entropy label multisets

20. 20 What Not to Do: Use Entropy Entropy: Improves at first, hurts in long run

21. Why not Entropy In the presence of noise, entropy will be high even with many labels Entropy is scale invariant {3+, 2-} has same entropy as {600+, 400-}, i.e., 0.97 21

22. 22 Estimating Label Uncertainty (LU) Observe +’s and –’s and compute Pr{+|obs} and Pr{-|obs} Use Bayesian estimate of Bernoulli: Beta prior + update Uncertainty = tail of beta distribution S LU 0.5 0.01.0 Beta probability density function

23. Label Uncertainty p=0.7 5 labels (3+, 2-) Entropy ~ 0.97 CDF  =0.34 23

24. Label Uncertainty p=0.7 10 labels (7+, 3-) Entropy ~ 0.88 CDF  =0.11 24

25. Label Uncertainty p=0.7 20 labels (14+, 6-) Entropy ~ 0.88 CDF  =0.04 25

26. Label Uncertainty vs. Round Robin 26 similar results across a dozen data sets Remember: GRR already better than single labeling

27. Why LU is a hack?

28. 28 More sophisticated label uncertainty Observe +’s and –’s and compute Pr{+|obs} and Pr{-|obs} Estimated Beta distribution = quality of labelers for the example Estimate the prior Pr(+) by iterating

31. More sophisticated LU improves labeling quality under class imbalance and fixes some pesky LU learning curve glitches 31 Both techniques perform essentially optimally with balanced classes

32. 32 Another strategy : Model Uncertainty (MU) Learning models of the data provides an alternative source of information about label certainty – (a random forest for the results to come) Model uncertainty: get more labels for instances that cause model uncertainty Intuition? – for modeling: why improve training data quality if model already is certain there? – for data quality, low-certainty “regions” may be due to incorrect labeling of corresponding instances Models Examples Self-healing process [Brodley et al, 99] + + + + + + + + + + - - - - - - - - - - - - - - - - +

33. 33 Yet another strategy : Label & Model Uncertainty (LMU) Label and model uncertainty (LMU): avoid examples where either strategy is certain

34. Label Quality 34 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 0400800120016002000 Number of labels (waveform, p=0.6) Labeling quality UNFMU LULMU Label Uncertainty Uniform, round robin Label + Model Uncertainty Model Uncertainty alone also improves quality

35. 35 Model Quality Label & Model Uncertainty Across 12 domains, LMU is always better than GRR. LMU is statistically significantly better than LU and MU.

36. Adult content classification 36

37. Example: Build an “Adult Web Site” Classifier Need a large number of hand-labeled sites Get people to look at sites and classify them as: G (general audience) PG (parental guidance) R (restricted) X (porn) Cost/Speed Statistics  Undergrad intern: 200 websites/hr, cost: $15/hr  MTurk: 2500 websites/hr, cost: $12/hr Cost/Speed Statistics  Undergrad intern: 200 websites/hr, cost: $15/hr  MTurk: 2500 websites/hr, cost: $12/hr

38. Bad news: Spammers! Worker ATAMRO447HWJQ labeled X (porn) sites as G (general audience) Worker ATAMRO447HWJQ labeled X (porn) sites as G (general audience)

39. Redundant votes, infer quality Look at our spammer friend ATAMRO447HWJQ together with other 9 workers  Using redundancy, we can compute error rates for each worker

40. 1.Initialize“correct” label for each object (e.g., use majority vote) 2.Estimate error rates for workers (using “correct” labels) 3.Estimate “correct” labels (using error rates, weight worker votes according to quality) 4.Go to Step 2 and iterate until convergence 1.Initialize“correct” label for each object (e.g., use majority vote) 2.Estimate error rates for workers (using “correct” labels) 3.Estimate “correct” labels (using error rates, weight worker votes according to quality) 4.Go to Step 2 and iterate until convergence Repeated Labeling, EM, and Confusion Matrices (Dawid & Skene, 1979) Iterative process to estimate worker error rates Our friend ATAMRO447HWJQ marked almost all sites as G. Seems like a spammer… Error rates for ATAMRO447HWJQ P[G → G]=99.947% P[G → X]=0.053% P[X → G]=99.153% P[X → X]=0.847%

41. Challenge: From Confusion Matrixes to Quality Scores The algorithm generates “confusion matrixes” for workers How to check if a worker is a spammer using the confusion matrix? (hint: error rate not enough) How to check if a worker is a spammer using the confusion matrix? (hint: error rate not enough) Error rates for ATAMRO447HWJQ  P[X → X]=0.847%P[X → G]=99.153%  P[G → X]=0.053%P[G → G]=99.947%

42. Challenge 1: Spammers are lazy and smart! Confusion matrix for spammer  P[X → X]=0% P[X → G]=100%  P[G → X]=0% P[G → G]=100% Confusion matrix for good worker  P[X → X]=80%P[X → G]=20%  P[G → X]=20%P[G → G]=80% Spammers figure out how to fly under the radar… In reality, we have 85% G sites and 15% X sites Errors of spammer = 0% * 85% + 100% * 15% = 15% Error rate of good worker = 85% * 20% + 85% * 20% = 20% False negatives: Spam workers pass as legitimate

43. Challenge 2: Humans are biased! Error rates for CEO of AdSafe P[G → G]=20.0%P[G → P]=80.0%P[G → R]=0.0%P[G → X]=0.0% P[P → G]=0.0%P[P → P]=0.0%P[P → R]=100.0%P[P → X]=0.0% P[R → G]=0.0%P[R → P]=0.0%P[R → R]=100.0%P[R → X]=0.0% P[X → G]=0.0%P[X → P]=0.0%P[X → R]=0.0%P[X → X]=100.0% Error rates for CEO of AdSafe P[G → G]=20.0%P[G → P]=80.0%P[G → R]=0.0%P[G → X]=0.0% P[P → G]=0.0%P[P → P]=0.0%P[P → R]=100.0%P[P → X]=0.0% P[R → G]=0.0%P[R → P]=0.0%P[R → R]=100.0%P[R → X]=0.0% P[X → G]=0.0%P[X → P]=0.0%P[X → R]=0.0%P[X → X]=100.0%  In reality, we have 85% G sites, 5% P sites, 5% R sites, 5% X sites  Errors of spammer (all in G) = 0% * 85% + 100% * 15% = 15%  Error rate of biased worker = 80% * 85% + 100% * 5% = 73% False positives: Legitimate workers appear to be spammers

44. Solution: Reverse errors first, compute error rate afterwards When biased worker says G, it is 100% G When biased worker says P, it is 100% G When biased worker says R, it is 50% P, 50% R When biased worker says X, it is 100% X Small ambiguity for “R-rated” votes but other than that, fine! Error Rates for biased worker P[G → G]=20.0%P[G → P]=80.0%P[G → R]=0.0%P[G → X]=0.0% P[P → G]=0.0%P[P → P]=0.0%P[P → R]=100.0%P[P → X]=0.0% P[R → G]=0.0%P[R → P]=0.0%P[R → R]=100.0%P[R → X]=0.0% P[X → G]=0.0%P[X → P]=0.0%P[X → R]=0.0%P[X → X]=100.0% Error Rates for biased worker P[G → G]=20.0%P[G → P]=80.0%P[G → R]=0.0%P[G → X]=0.0% P[P → G]=0.0%P[P → P]=0.0%P[P → R]=100.0%P[P → X]=0.0% P[R → G]=0.0%P[R → P]=0.0%P[R → R]=100.0%P[R → X]=0.0% P[X → G]=0.0%P[X → P]=0.0%P[X → R]=0.0%P[X → X]=100.0%

45. When spammer says G, it is 25% G, 25% P, 25% R, 25% X When spammer says P, it is 25% G, 25% P, 25% R, 25% X When spammer says R, it is 25% G, 25% P, 25% R, 25% X When spammer says X, it is 25% G, 25% P, 25% R, 25% X [note: assume equal priors] The results are highly ambiguous. No information provided! Error Rates for spammer: ATAMRO447HWJQ P[G → G]=100.0%P[G → P]=0.0% P[G → R]=0.0%P[G → X]=0.0% P[P → G]=100.0% P[P → P]=0.0% P[P → R]=0.0% P[P → X]=0.0% P[R → G]=100.0% P[R → P]=0.0%P[R → R]=0.0% P[R → X]=0.0% P[X → G]=100.0% P[X → P]=0.0%P[X → R]=0.0%P[X → X]=0.0% Error Rates for spammer: ATAMRO447HWJQ P[G → G]=100.0%P[G → P]=0.0% P[G → R]=0.0%P[G → X]=0.0% P[P → G]=100.0% P[P → P]=0.0% P[P → R]=0.0% P[P → X]=0.0% P[R → G]=100.0% P[R → P]=0.0%P[R → R]=0.0% P[R → X]=0.0% P[X → G]=100.0% P[X → P]=0.0%P[X → R]=0.0%P[X → X]=0.0% Solution: Reverse errors first, compute error rate afterwards

46.  Cost of “soft” label with cost c ij for labeling as j an object of class i  “Soft” labels with probability mass in a single class are good  “Soft” labels with probability mass spread across classes are bad Cost (spammer) = Replace p i with Prior i Computing Quality Scores Quality = 1 – Cost(worker) / Cost(spammer)

47. Experimental Results 500 web pages in G, P, R, X 100 workers per page (just to evaluate effect of more labels) 339 workers Lots of noise! 95% accuracy with majority vote (only!) Dropped all workers with quality score 50% or below Error rate: 1% of labels dropped Quality scores: 30% of labels dropped, accuracy 99.8% Note massive amount of redundancy and very conservative spam rejection

48. Too much theory? Open source implementation available at: http://code.google.com/p/get-another-label/ Input: – Labels from Mechanical Turk – Cost of incorrect labelings (e.g., X  G costlier than G  X) Output: – Corrected labels – Worker error rates – Ranking of workers according to their quality Beta version, more improvements to come! Suggestions and collaborations welcomed!

49. Workers reacting to quality scores Score-based feedback leads to strange interactions: The angry, has-been-burnt-too-many-times worker: “F*** YOU! I am doing everything correctly and you know it! Stop trying to reject me with your stupid ‘scores’!” The overachiever: “What am I doing wrong?? My score is 92% and I want to have 100%” 49

50. An unexpected connection at the NAS “Frontiers of Science” conf. 50 Your spammers behave like my mice!

51. An unexpected connection at the NAS “Frontiers of Science” conf. 51 Your spammers behave like my mice! Eh?

52. An unexpected connection at the NAS “Frontiers of Science” conf. 52 Your spammers want to engage their brain only for motor skills and not for cognitive skills Yeah, makes sense…

53. An unexpected connection at the NAS “Frontiers of Science” conf. 53 And here is how I train my mice to behave… I should try this the moment that I get back to my room

54. Implicit Feedback with Frustration Punish bad answers with frustration (e.g, delays between tasks) – “Loading image, please wait…” – “Image did not load, press here to reload” – “Network error. Return the HIT and accept again” Reward good answers by giving next task immediately, with no problems →Make this probabilistic to keep feedback implicit 54

55. First result Spammer workers quickly abandon Good workers keep labeling Bad: Spammer bots unaffected How to frustrate a bot? – Give it a CAPTHCA 55

56. Second result (more impressive) Remember, Don was training the mice… 15% of the spammers start submitting good work! Putting cognitive effort is more beneficial … Key trick: Learn to test workers on-the-fly – Model performance using Dirichlet compound multinomial distribution (details offline) 56

57. Thanks! Q & A?

58. Why does Model Uncertainty (MU) work? MU score distributions for correctly labeled (blue) and incorrectly labeled (purple) cases 58 + + + + + + + + + + - - - - - - - - - - - - - - - - +

59. 59 Why does Model Uncertainty (MU) work? Models Examples Self-healing process+ + + + + + + + + + - - - - - - - - - - - - - - - - + Self-healing MU “active learning” MU

60. What if different labelers have different qualities? (Sometimes) quality of multiple noisy labelers is better than quality of best labeler in set here, 3 labelers: p-d, p, p+d 60

61. Soft Labeling vs. Majority Voting MV: majority voting ME: soft labeling

62. Related topic Estimating (and using) the labeler quality – for multilabeled data: Dawid & Skene 1979; Raykar et al. JMLR 2010; Donmez et al. KDD09 – for single-labeled data with variable-noise labelers: Donmez & Carbonell 2008; Dekel & Shamir 2009a,b – to eliminate/down-weight poor labelers: Dekel & Shamir, Donmez et al.; Raykar et al. (implicitly) – and correct labeler biases: Ipeirotis et al. HCOMP-10 – Example-conditional labeler performance Yan et al. 2010a,b Using learned model to find bad labelers/labels: Brodley & Friedl 1999; Dekel & Shamir, Us (I’ll discuss) 62