Bibliographic record
Abstract
Many machine learning algorithms learn from the data by capturing certain interesting characteristics. Decision trees are used in many classification tasks. In some circumstances, we only want to consider fixed-depth trees. Unfortunately, finding the optimal depth-d decision tree can require time exponential in d. In the first part of this dissertation, we present OPTNBDT algorithm, which is a fast way to produce a near optimal fixed-depth decision tree under the Naıve Bayes assumption. We apply this technique to real-world datasets and find that our model improves the computational costs significantly while yielding relatively high accuracy. In the second part of this dissertation, we present two novel algorithms for active learning. There are scenarios where we have access to many unlabeled data; however, obtaining the labels for the data is difficult. An active learning process tries to address this issue by selectively requesting the label of few unlabeled instances, with the goal of using these newly labeled instances to produce an effective classifier. First, we focus on active learning for image segmentation, which requires producing a label for each pixel in an image. We provide an active learner (LMU) that first selects the image whose expected label will reduce the uncertainty of other unlabeled images the most, and then after greedily selects the most informative image. The results of our experiments, over real-world datasets show that training on very few informative images can produce a segmenter that is as good as training on the entire dataset. Finally, we present the importance sampling algorithm (ISAL) for actively learning in the standard classification framework, which we demonstrate its sample and label efficiency. In particular, on each iteration, ISAL identifies a distribution that puts large weight on instances whose labels are uncertain, then requests the label of an instance drawn from that distribution. We prove that ISAL can be more sample and label efficient than passive learning with an exponential convergence rate to the Bayes classifier on noise-free data. We also provide empirical studies that show ISAL is more effective than many other active learning algorithms.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.017 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.005 | 0.006 |
| Open science | 0.004 | 0.004 |
| Research integrity | 0.003 | 0.005 |
| Insufficient payload (model declined to judge) | 0.011 | 0.007 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".