Robustness properties of some bayesian inferences
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
In this thesis, we investigate the local robustness of traditional Bayesian inferences and the robustness of observed relative surprise (ORS) inferences, proposed in Evans (1997), under the contamination model and within the parametric family. Since the concept of observed surprise (OS) is corresponds to the traditional Bayesian inferences, the Gateaux derivatives for OS and ORS, under two situations, are firstly calculated and compared. Then robustness of inferences such as estimation and hypothesis testing via two methods are investigated. From the derived closed form of Gateaux derivatives, it is easy to see that ORS inferences are always robust, in the sense that they have bounded derivatives and, mostly are more robust than traditional Bayesian inferences, in the sense that their absolute value of Gateaux derivatives are smaller than that of traditional Bayesian inferences. For traditional Bayesian inferences themselves, these Gateaux derivative formulae show that the robustness is affected by the geometrical shape of the posterior density functions: when posteriors take the shape of unimodal (or multimodal) or are U-shaped, the derivative could be unbounded, but when it is monotone the derivative is bounded and similar to that of ORS inferences.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
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 it