branch_selection

"heterogeneous per-dimension priors reduce latent permutation symmetry relative to homogeneous shared priors... KL regularization induced by the adaptive GMM prior creates source-specific attraction behavior"

"MGCE margin loss is classification-calibrated... arg max f(x,μ*)y = arg max p*(y|x)"

"both AIMER and the Hoyer metric are functions of the same underlying ℓ1/ℓ2 ratio"

"By examining the action of the supercharge deformed by an exactly marginal operator on the relevant symmetry algebra, we propose a superselection rule governing the lifting process of BPS states, and use it to construct the REG by summing only over those symmetry sectors that can mix according to this rule."

"η_TLQLC(s) and η_TLAH(s) efficiency curves... ARE bounded 0.95–1.01"

"comparative risk-aware redesign based on right-tail Conditional Value-at-Risk (RU-CVaR)"

"Kriging Believer variance deflation σ^(j)²(x) = σ^(0)²(x) − Σ k(x,x^(l))² / (σ²(x^(l)) + σ_n²)"

"four-dimensional grid search over cluster count (K from 3 to 20), latent dimensionality, network depth, and channel width"

"WNE is parameter-free; it exactly parameterizes the NE class without constraining internal layers."

"Proposition 1.8 and Theorem 1.6: first-moment + branching-number lower bound θ_c ≥ 1/(e br_T)"

"The priority function: w_i,f(t) = 1/5QI × r_i,f(t)/R_i,f(t)"

"Tensor interface exports per-token entropy H_t = −∑ p_v^{(t)} log p_v^{(t)} that is structurally coupled to correctness under standard training."

"prioritizes labeling high-utility data while balancing the tradeoff between training and labeling cost"

"We trace this gain to markedly straighter probability paths, echoing the effect of optimal-transport couplings in Rectified Flow"

"From a single quadratic cost, joint minimization yields a unified solution that decomposes orthogonally into two distinct channels. The pace channel compresses execution along the planned direction, while the path channel applies an orthogonal spatial offset"

"softales ... enriched poset that generalises ... residuated lattices ... featuring a family of operations internalizing ⊕_{-p} and ⊕_p via an enriched universal property"

"data scaling laws, cross-domain artifacts transfer-interference patterns"

"we propose a bilinear network-aware tokenization... W_{l,m}=U_l ⊙ U_m ... replaces quadratic growth in patch-specific parameters with a linear scaling in the number of networks"

"We construct three independent low-rank bilinear modules: (h×r),(h×c),(r×c). Each module learns multiplicative relationships using a factorized bilinear mapping: f(x, y) = W((U x) ⊙ (V y))"

"POMA formulates knowledge transfer as retrieve–compose–adapt with GRPO policy optimization over candidate pool C, dual-scale alignment LDA using squared Frobenius norms of covariance matrices at mol and sub scales."

"We observe an enhanced visibility of charge-density-wave correlations... after a rapid boost of the interaction strength"

"q*_i = exp(η g_i(θ)) / sum exp(η g_j(θ))"

"WD+MC fits to TESS LCs … hot spot model … fill-out factor 0.27"

"RoSHAP_j := (1−P0j) m_j² / s_j ... rewards features that are active, strong, and stable."

"vGDD := max_{g_i(·)} ∑ min (f_i(x_i)+g_i(x_i)) s.t. ∑ g_i(x)=0 ∀x∈X"

"contrastive hinge loss... |p_i - p_j|^2 for true pairs, max(0, m - |p_i - p_j|^2) otherwise"

"Define ϕ^-_M(t) := max{0, a-M t, b-M(1-t)} ... Φ^-(M) ..."

"random group model μ_Γ on admissible Γ-groups with H-moments |H^Γ| ∏_{v∈T} |H^{Γ_v}|"

"Empirical evaluations demonstrate state-of-the-art performance on imbalanced datasets under two labeling mechanisms - selecting positives completely at random (SCAR) and selecting at random (SAR)."

"Flow Matching regression LCFM = E ∥vθ(zt,t) − (z1 − z0)∥² as Lagrangian surrogate for HJB"