NG21A-05 Real-time resolution of instrumental biases using Rao-Blackwellized Particle Filtering
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
Particle filters are a non-linear data assimilation technique which use an ensemble of states to approximate the posterior density of the modelled geophysical system. While particle filters have nearly unparalleled flexibility to handle non-linear and non-Gaussian measurements, states, and errors, this comes at a cost of require comparatively large ensemble sizes when compared to other ensemble techniques. When the instruments which provide data to the model are subject to biases or calibration errors, each additional bias to be corrected adds another dimension to the state space, reducing the sampling efficiency of the entire ensemble. As the maximum ensemble size is generally limited by computational cost, this can create the perverse situation where adding additional observations decreases the performance of the data assimilation. Rao-Blackwellized particle filtering presents an efficient solution to this problem, allowing for the analytical solution of a conditionally linear Gaussian subset of the state space with minimal computational cost. This technique has been demonstrated successfully in two operational particle filter data assimilation models, the regional Assimilative Canadian High Arctic Ionospheric Model (A-CHAIM) and the global, real-time ionosphere/plasmasphere model Advanced Ionospheric Data Assimilation (AIDA). These models rely on the thousands of Global Navigation Satellite System (GNSS) receivers which provide an integrated measurement of electron density, but are subject to time-varying hardware-specific biases. The biases produced by these models will be compared to independently-derived values, and the improvement in model performance will be demonstrated.
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 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 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