File size: 6,993 Bytes
19591f3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
{"claim_id": "planar_profile", "details_json": "{\"origin\": \"Rank-one result is established prior work; the supplied predecessor proves the planar all-budget extension, reproduced here.\"}", "qualification": "", "source_path": "research/RESEARCH_STATUS.json", "statement": "F(r,n)=2n/(2n-1) for affine ranks r=1,2 and every n>=1.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "rare_simplex_exact_family", "details_json": "{\"status\": \"Analytic proof with complete count-pattern reduction.\"}", "qualification": "", "source_path": "research/RESEARCH_STATUS.json", "statement": "Exact optimal selection error for anchor mass 1-t and k equally weighted rare simplex vertices, including budgets exceeding k.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "all_budget_obstruction", "details_json": "{}", "qualification": "Lower bound for unrestricted profile; finite-dataset supremum justified by rational approximation when optimal masses are irrational.", "source_path": "research/RESEARCH_STATUS.json", "statement": "F(k,n)>=1+B(k,n) for every k>=4,n>=2, with B=[2n(k-2)-(k-3)-sqrt((k-3)(2n(n-2)(k-2)+(k-3)))]/(k*n^2).", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "strict_dimension_four_separation", "details_json": "{\"consequence\": \"The planar law fails at every nontrivial budget in every dimension at least four.\"}", "qualification": "", "source_path": "research/RESEARCH_STATUS.json", "statement": "B(4,n)>1/(2n-1) for every n>=2.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "rational_coordinate_witness", "details_json": "{\"verification_report\": \"obstruction_verification.json\"}", "qualification": "", "source_path": "research/RESEARCH_STATUS.json", "statement": "A 128-example integer-coordinate dataset in R^4 has exact budget-three full loss ratio 13810/11457>6/5.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "asymptotic_obstruction", "details_json": "{}", "qualification": "Strictly above 1/2 for k>4; not asserted to be the complete asymptotic optimum.", "source_path": "research/RESEARCH_STATUS.json", "statement": "liminf_n n(F(k,n)-1)>=[2(k-2)-sqrt(2(k-2)(k-3))]/k for k>=4.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "all_rank_constructive_upper", "details_json": "{}", "qualification": "Not claimed sharp for r>=3,n>=3.", "source_path": "research/RESEARCH_STATUS.json", "statement": "The explicit pair-seeded and propagated coefficient U_r(n) bounds every affine-rank-r dataset; U_r(n)<1/n for finite r and n>=2.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "residual_allocation", "details_json": "{}", "qualification": "Fixed-geometry asymptotics do not determine the budget-dependent worst-case profile.", "source_path": "research/RESEARCH_STATUS.json", "statement": "Floor multiplicities plus fractional-residual selection give an additional exact instance certificate and O(n^-2) squared error for fixed finite support.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "symmetric_profile_minimax", "details_json": "{}", "qualification": "Exact fixed-profile value; does not optimize over arbitrary probability vectors.", "source_path": "research/RESEARCH_STATUS.json", "statement": "For every k>=2,n>=1,0<t<1, the geometric minimax at masses (1-t,t/k,...,t/k) equals a finite two-mode game with an optimal dual supported on at most two balanced rare-count types.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "closed_form_matching_profile", "details_json": "{}", "qualification": "", "source_path": "research/RESEARCH_STATUS.json", "statement": "For k>=4 and n>=2, at the explicitly optimized mass t*, the symmetric-profile minimax equals B(k,n); a one-rare/two-rare count lottery has expected error exactly B(k,n)V for every support geometry.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "residue_pruning", "details_json": "{\"algorithm_cost\": \"O(min(k,n+1)^3) exact scalar arithmetic for the unrestricted symmetric-profile value, excluding bit costs; explicit allowed lists add O(|A|) reading cost.\"}", "qualification": "", "source_path": "research/RESEARCH_STATUS.json", "statement": "Within each residue modulo k, retain only a legal rare-count total nearest nt; at most min(k,n+1) candidates remain.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "arbitrary_rule_minimax", "details_json": "{}", "qualification": "Smaller fixed dimensions impose an additional nonconvex rank constraint; unrestricted dual remains only an upper bound there.", "source_path": "research/RESEARCH_STATUS.json", "statement": "Any nonempty finite admissible integer count set has exact primal normalized-Gram and dual count-covariance formulations when ambient dimension up to m-1 is allowed.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "matrix_and_margin_rules", "details_json": "{}", "qualification": "", "source_path": "research/RESEARCH_STATUS.json", "statement": "A universal scalar selection constant lifts to one randomized policy with a matrix bound in expectation for all fixed quadratic metrics; an explicit positive-margin condition gives deterministic linear-rule feasibility with quantified objective penalty.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "unbiasedness_barrier", "details_json": "{}", "qualification": "Mean unbiasedness is a substantive additional requirement.", "source_path": "research/RESEARCH_STATUS.json", "statement": "The best universal expected squared-error factor for mean-unbiased equal-weight n-draw selectors is exactly 1/n.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "quadratic_erm", "details_json": "{}", "qualification": "The full Hessian is positive definite; c is valid for the lifted curvature-residual feature rank; the full Hessian, optimizer and residual features must be available.", "source_path": "research/RESEARCH_STATUS.json", "statement": "For varying-Hessian convex quadratic losses, if cM<1 then a selected ERM has excess at most cG/[2(1-sqrt(cM))^2].", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "replay_reuse_and_recursive_composition", "details_json": "{}", "qualification": "Adaptive reference process uses the same realized rule sequence; indefinite stability requires the stated contraction or error-summability conditions.", "source_path": "research/RESEARCH_STATUS.json", "statement": "Gradient and higher-derivative certificates control reuse radii; observable error bounds certify descent; pathwise local errors compose through explicit Lipschitz products.", "status": "stated_and_proved_in_manuscript"}
{"claim_id": "population_transfer", "details_json": "{}", "qualification": "No claim that selected observations remain independent.", "source_path": "research/RESEARCH_STATUS.json", "statement": "Under iid finite-variance sampling, certified empirical mean error yields expected population excess <=sigma^2(sqrt(c(N-1)/N)+1/sqrt(N))^2.", "status": "stated_and_proved_in_manuscript"}