Bibliographic record
Abstract
We have entered an era when most data has been outsourced to a potentially untrusted server. In this setting, even with encryption, the physical location accessed, or access patterns, during a query can reveal significant information about the query or database. This motivates the development of techniques to conceal access patterns, such as Private Information Retrieval (PIR) and Oblivious RAM (ORAM). Both allow retrieval of database item without revealing which item is retrieved at a sub-linear (often polylog) communication/computation cost to the client. ORAM requires the client to have a secret key and the server to keep state, but gets sub-linear (even log) server computation and allows both reads and writes. PIR does not require any client secret and the server is stateless, but requires linear server computation. This motivated the development of Doubly Efficient Private Information Retrieval (DEPIR) by Canetti et al. [CHR'17] and Boyle et al. [BIPW'17]. DEPIR allows a client to interact with a single stateless server and achieve overhead (both bandwidth and server computation) that is sub-linear in the database size. In this thesis, we explore three applications of DEPIR in the outsourced data setting, and tradeoffs that lie between ORAM and PIR. We leverage DEPIR's inherent statelessness and sub-linearity to achieve applications that neither PIR nor ORAM could accomplish. First we will explore a variant of ORAM, Rewindable ORAM (RORAM) that leverages DEPIR's statelessness to maintain security even when an adversary resets client and server state. RORAM is required to create the first heuristic construction of Fully Homomorphic Encryption in the RAM model. Our second application considers secure computation in the RAM model, using DEPIR's sub-linear server work to construct a distributed ORAM with constant rounds and sub-linear server work. Finally, we use techniques from DEPIR to achieve a multi-client private and anonymous data access scheme called PANDA.--Author's abstract
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 | 0.002 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.006 |
| Open science | 0.002 | 0.004 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.005 | 0.002 |
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".