Topics — What a work is about. Nucleation holds whatever reading the owner files here, whether or not an Anthology of the SOTA topic could also hold it; instruction for ML practice belongs to the anthology (ADR-048). The words are seeded from the clusters the reading-time triage of 2026-09-25 found, and grow by decision.
Compositionality — how the structure and meaning of a whole come from its parts and how they are combined — the principle of compositionality and compositional semantics, the emergence of compositional structure in languages under transmission, and composition and binding in learned models (generalization to unseen combinations, attribute binding, composing distributions or concepts). A work on making a model generalize compositionally is ML practice and goes to the anthology (ADR-031).
9 of 371 QUESTION documents. Back to the full index.
| # | Title | Summary | Status |
|---|---|---|---|
| QUESTION-012 | When does reconstruction through an intermediate image preserve or degrade observational structure relative to direct text-to-text reconstruction? | Proposal v6’s Result 5 (A100), “especially interesting as a proposition that connects the formal theory to a novel empirical setting”. Dropped from the manuscript, which keeps the text→image→caption chain as Case III but states no result about composition through a medium. | Open |
| QUESTION-018 | When a generator receives several conditions, what determines whether it preserves their conjunction and relational bindings? | The fourth of the five central questions (A110 §IX). The manuscript states the failure (§7) and its benchmarks (§10) but not the question. | Open |
| QUESTION-026 | Does categorical compositional semantics constrain transport between context-indexed empirical models, so that how a whole utterance’s observables are carried is fixed by how its parts’ are, and does the constraint reach past propositional content? | Raised in a review of the record that the owner relayed on 2026-10-09, which asked for a direct comparison between DisCoCat (LIT-273, LIT-272) and the manuscript’s transport. The record already holds the meeting point the review did not mention: Abramsky and Sadrzadeh’s Semantic Unification (LIT-841). The sharp edge is the manuscript’s own: what DisCoCat composes is propositional and distributional meaning, which CLAIM-061 says is not enough to preserve the act. | Open |
| QUESTION-039 v2 | What sets the order and the moment, in training time or in examples, at which a network acquires a structure, and is acquisition one kind of event or several? | Filed in the fourth synthesis pass to give readings no question asked for a place in the argument. It bears on CLAIM-164. THEORY-086, THEORY-187, THEORY-200, THEORY-204 and THEORY-039 answer it in part; the body gives each one’s limit. | Open |
| QUESTION-165 | On the Random Hierarchy Model, is the number of examples a deep network needs set by when patch-label correlations become detectable, or by the size of the grammar, once the two are separated by moving the correlations with n_c, m and L fixed? | The promotion condition formerly on THEORY-195, moved here unchanged when THEORY-195 was narrowed and promoted on 2026-10-19 to what the experiments measured. In the model as given, the correlation threshold and the grammar’s size are the same expression, n_c m^L, so the measurements cannot say which one sets the number of examples. | Open |
| QUESTION-178 | Is word co-occurrence close to independent, multiplicative binary attributes in the analogy families where spectral embeddings complete parallelograms, and does adding an interaction between two attributes break their parallelogram? | Filed when the promotion audit (batch 037, ADR-047 v2) narrowed and promoted THEORY-185, Korchinski et al.’s conditional theorem: under independent, multiplicative binary attributes the PMI is affine in the attributes and its spectral embedding gives exact parallelograms. The theorem is conditional on that premise. Its former promote_when asked for the premise to be measured on real text, and this question holds that request. | Open |
| QUESTION-207 | Without choosing checkpoints on the unseen pairs, does coverage of combinations still beat volume for compositional generalization, in a second architecture or with more than two concepts? | The promotion condition formerly on THEORY-117. THEORY-117 was narrowed to what Uselis, Dittadi and Oh measured and promoted on 2026-10-20 (ADR-047 v3), with a determination result the record re-derived: from-scratch ResNet-50s on two-concept grids, with checkpoints chosen on the unseen pairs, reach unseen combinations as coverage grows and not under up to four times more data of the same combinations. Whether that survives honest model selection, another architecture or more concepts is open. | Open |
| QUESTION-283 | Can composing systems create an effective structure whose adequate specification needs a richer relational vocabulary than the parts needed separately (relations of higher arity, history, or a new kind) and, if so, under what conditions, and is the new vocabulary forced by the composite or only by the restricted class of descriptions in use? | Raised at the owner’s review of the pull request, which called it the result most worth preserving and asked that it “become a major research question”. Hansen and Ghrist show that eliminating an interior vertex of a cellular sheaf can leave a three-way boundary relation, v₁ = v₂ + v₃, that no sheaf on the boundary’s pairwise edges expresses (THEORY-1090; CLAIM-480). For scalar networks reduction stays pairwise. The question asks when composition forces a change in the type of relational structure an adequate description needs. It also asks whether that is a fact about the composite or about the class of descriptions, and whether the same thing is what happens when a record’s relation scheme cannot express what a composed organization requires. | Open |
| QUESTION-343 | Does a rerun of Sarkar, Mukherjee and Sato’s released harness, or an independent implementation, reproduce Table 1’s single-path, uniform and holonomy-weighted accuracies on S4 synchronization with outliers within their intervals, including the paired holonomy − uniform gains? | The promotion condition formerly on THEORY-1158, moved here when the status review of 2026-10-30 promoted it under ADR-065. The accuracies are the paper’s, on 1000 instances per cell, and were not rerun; the AUC ceiling, which is arithmetic, was recomputed and needs no rerun. | Open |