King Mixture Representations

• King Mixture Representations are finite-mixture reparametrizations of proper species sampling processes that preserve the original infinite-dimensional law through a latent truncation variable and reweighted atoms.
• They maintain exchangeability and the full predictive structure by converting an infinite paintbox into a finite mixture framework, facilitating practical MCMC implementations.
• This approach connects to broader mixture theory, with analogues in Gaussian mixture embeddings and King function approximations, thereby broadening its applications in Bayesian nonparametrics and related fields.

to=arxiv_search.^^^^^^^^2query2^^^^^^^^ 天天乐购彩票json {"^^^^^^^^2query2 OR \2^^^^^^^^"Exact finite mixture representations for species sampling processes\"","max_results":5,"sort_by":"submittedDate","sort_order":"descending"}【อ่านข้อความเต็มjson to=arxiv_search.^^^^^^^^2query2^^^^^^^^ েচ্ছিন্ন qq的天天中彩票 [{"arxiv_id":"(&&&^^^^2query2^^^^&&&)","title":"Exact finite mixture representations for species sampling processes","authors":["Salvatore Favaro","Vincenzo Vitale"],"summary":"Random probability measures, together with their constructions, representations, and associated algorithms, play a central role in modern Bayesian inference. A key class is that of proper species sampling processes, which offer a relatively simple yet versatile framework that extends naturally to non-exchangeable settings. We revisit this class from a computational perspective and show that they admit exact finite mixture representations. In particular, we prove that any proper species sampling process can be written, at the prior level, as a finite mixture with a latent truncation variable and reweighted atoms, while preserving its distributional features exactly. These finite formulations can be used as drop-in replacements in Bayesian mixture models, recasting posterior computation in terms of familiar finite-mixture machinery. This yields straightforward MCMC implementations and tractable expressions, while avoiding ad hoc truncations and model-specific constructions. The resulting representation preserves the full generality of the original infinite-dimensional priors while enabling practical gains in algorithm design and implementation.","categories":["stat.ME","math.PR"],"published":"2^^^^2query22^^^^5-^^^^2id:(Mena et al., 30 Dec 2025) OR \22^^^^-3^^^^2query2^^^^","pdf_url":"http://arxiv.org/pdf/(&&&^^^^2query2^^^^&&&)v^^^^2id:([2512.24414](/papers/2512.24414)) OR \2^^^^","abs_url":"http://arxiv.org/abs/(&&&^^^^2query2^^^^&&&)v^^^^2id:([2512.24414](/papers/2512.24414)) OR \2^^^^"}] to=arxiv_search.^^^^^^^^2query2^^^^^^^^ 天天彩աիjson {"^^^^^^^^2query2 OR id:(Baringhaus et al., 2022) OR id:(Wang et al., 5 Jun 2026)","max_results":^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^,"sort_by":"submittedDate","sort_order":"descending"} 开元棋牌 to=arxiv_search.^^^^^^^^2query2^^^^^^^^ 天天中彩票公司 񹚂json [{"arxiv_id":"(Wang et al., 5 Jun 2026)","title":"King Function for Shifted Gaussian: Laguerre Structure, Spectral Theory and Density","authors":["Guangyao Zhang"],"summary":"We study King function arising as radial kernels in the laboratory-frame spherical harmonic expansion of shifted Gaussian distributions. We first clarify their relation with the co-moving Laguerre hierarchy by means of a King-Laguerre expansion. We then derive the King differential equation and show that the associated self-adjoint operator in a Gaussian-weighted Hilbert space is unitarily equivalent to the free radial Schrodinger operator on the half-line. This yields the spectral representation and generalized eigenfunction. Finally, we prove that real-parameter King function, lies in the resolvent set, form a dense non-orthogonal system in a natural radial velocity space, providing an approximation-theoretic basis for King mixture representations. Weighted L^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^-integrability criteria and closed-form moment formulas are also derived, justifying the normalization of King function.","categories":["math-ph","math.SP","physics.plasm-ph"],"published":"2^^^^2query22^^^^6-^^^^2query2 OR \2^^^^","abs_url":"http://arxiv.org/abs/([2606.12455](/papers/2606.12455))v^^^^2id:([2512.24414](/papers/2512.24414)) OR \2^^^^"},{"arxiv_id":"(Baringhaus et al., 2022)","title":"Discrete mixture representations of parametric distribution families: geometry and statistics","authors":["Lutz Mattner","Stephan Schrempp"],"summary":"We investigate existence and properties of discrete mixture representations PRESERVED_PLACEHOLDER_^^^^2query2^^^^ for a given family PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^, $\theta\in\Theta$, of probability measures. The noncentral chi-squared distributions provide a classical example. We obtain existence results and results about geometric and statistical aspects of the problem, the latter including loss of Fisher information, Rao-Blackwellization, asymptotic efficiency and nonparametric maximum likelihood estimation of the mixing probabilities.","categories":["math.ST","math.PR","62F^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^, 62F^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2^^^^, 6^^^^2query2^^^^E^^^^2query2 62F^^^^2id:(Mena et al., 30 Dec 2025) OR \22^^^^"],"published":"2^^^^2query222^^^^- OR \2^^^^","abs_url":"http://arxiv.org/abs/([2206.11094](/papers/2206.11094))v^^^^2id:([2512.24414](/papers/2512.24414)) OR \2^^^^"},{"arxiv_id":"(Alaniz et al., 2022)","title":"Compositional Mixture Representations for Vision and Text","authors":["Sebastian D. A. Lippl","Jörn-Henrik Jacobsen","Janis Keuper"],"summary":"Learning a common representation space between vision and language allows deep networks to relate objects in the image to the corresponding semantic meaning. We present a model that learns a shared Gaussian mixture representation imposing the compositionality of the text onto the visual domain without having explicit location supervision. By combining the spatial transformer with a representation learning approach we learn to split images into separately encoded patches to associate visual and textual representations in an interpretable manner. On variations of MNIST and CIFAR^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^, our model is able to perform weakly supervised object detection and demonstrates its ability to extrapolate to unseen combination of objects.","categories":["cs.CV","cs.LG"],"published":"2^^^^2query222^^^^-^^^^2query2 OR \2^^^^3","pdf_url":"http://arxiv.org/pdf/([2206.06404](/papers/2206.06404))v^^^^2id:([2512.24414](/papers/2512.24414)) OR \2^^^^","abs_url":"http://arxiv.org/abs/([2206.06404](/papers/2206.06404))v^^^^2id:([2512.24414](/papers/2512.24414)) OR \2^^^^"}] “King mixture representations” is a nonstandard label applied to several distinct mixture-theoretic constructions. Its most precise usage in current Bayesian nonparametric work refers to the exact finite mixture representation of a proper species sampling process (SSP), obtained by introducing a latent truncation variable and reweighted atoms in a way that is compatible with Kingman’s paintbox representation of exchangeable partitions (&&&^^^^2query2^^^^&&&). The term is not standard in the general literature on discrete mixture representations (Baringhaus et al., 2022). Distinct, non-equivalent usages also appear in multimodal representation learning, where images and text are embedded as Gaussian mixtures in a shared latent space (Alaniz et al., 2022), and in mathematical physics, where mixtures of King functions are used to approximate shifted-Gaussian radial structures (Wang et al., 5 Jun 2026).

^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^. Terminological status and scope

In the Bayesian nonparametric setting, the phrase denotes an exact finite-mixture reparametrization of an infinite-dimensional prior. The supplied synthesis explicitly introduces the name “King mixture representations” to emphasize compatibility with Kingman’s representation of exchangeable partitions. In that sense, the object of study is not a new stochastic process class, but a finite-mixture representation of any proper SSP that preserves the original law exactly at the prior level (&&&^^^^2query2^^^^&&&).

In broader mixture theory, the phrase has no established canonical meaning. The paper on discrete mixture representations of parametric families states that “The term ‘King Mixture Representations’ is not standard in the mixture literature and does not appear in the paper.” Its focus is instead the general identity

$P_\theta=\sum_{i\in E} w_\theta(i)\,Q_i,$

together with existence, geometry, Fisher-information loss, Rao–Blackwellization, asymptotic efficiency, and NPMLE of the mixing probabilities (Baringhaus et al., 2022).

The phrase is therefore best understood as context-dependent. A concise classification is:

Context	Core object	Status of the term
Proper SSPs and exchangeable partitions	Exact finite mixture with latent truncation $K$	Explicitly introduced in the supplied synthesis
General discrete mixture theory	$P_\theta=\sum_i w_\theta(i)Q_i$	Not standard
Shifted-Gaussian radial analysis	Mixtures of King functions $\mathcal K_l$	Approximation-theoretic usage

This suggests that the SSP usage is the most technically specific sense of the term, while the others are analogical or domain-specific extensions.

2. Species sampling processes, exchangeability, and Kingman structure

Let $(X,\mathcal B_X)$ be a Polish space. A general SSP is a random probability measure

$$G(A)=\sum_{j=1}^\infty w_j\delta_{\theta_j}(A)+w_0G_0(A),\qquad A\in\mathcal B_X,$$

where $\theta_j$ are i.i.d. from a diffuse base measure PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^, the nonnegative weights satisfy PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2^^^^ with PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \22^^^^, and PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^3 is independent of PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^4. A proper SSP is the purely atomic case

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^5

so PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^6 and PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^7 has no residual diffuse component (&&&^^^^2query2^^^^&&&).

If PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^8 are i.i.d. from

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^9

the sample induces an exchangeable random partition $\theta\in\Theta$^^^^2query2^^^^ of $\theta\in\Theta$^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^ by tying observations that draw the same atom. The law of $\theta\in\Theta$2 is described by the exchangeable partition probability function (EPPF) $\theta\in\Theta$3, where $\theta\in\Theta$4 is the number of blocks and $\theta\in\Theta$5 are the block sizes. For classical Gibbs-type SSPs, including Dirichlet and Pitman–Yor priors, the EPPF admits closed form and determines predictive rules (&&&^^^^2query2^^^^&&&).

Kingman’s paintbox provides the structural backdrop. Kingman’s theorem states that any exchangeable random partition arises by sampling i.i.d. “colors” with frequencies $\theta\in\Theta$6 from a random paintbox. Proper SSPs realize such partitions with paintbox frequencies $\theta\in\Theta$7. In particular, Poisson–Dirichlet and Pitman–Yor priors are paintbox partitions with specific distributions on $\theta\in\Theta$8. In the SSP usage of the term, “King” refers to this Kingman compatibility rather than to a new probabilistic object (&&&^^^^2query2^^^^&&&).

3. Exact finite mixture representation

The central theorem for proper SSPs introduces a latent truncation variable $\theta\in\Theta$9 and a family of reweighted atoms. Let

$P_\theta=\sum_{i\in E} w_\theta(i)\,Q_i,$^^^^2query2^^^^

with $P_\theta=\sum_{i\in E} w_\theta(i)\,Q_i,$^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^ almost surely, and let $P_\theta=\sum_{i\in E} w_\theta(i)\,Q_i,$2 be a strictly decreasing sequence in $P_\theta=\sum_{i\in E} w_\theta(i)\,Q_i,$3 with $P_\theta=\sum_{i\in E} w_\theta(i)\,Q_i,$4. For $P_\theta=\sum_{i\in E} w_\theta(i)\,Q_i,$5, define

$P_\theta=\sum_{i\in E} w_\theta(i)\,Q_i,$6

Conditionally on $P_\theta=\sum_{i\in E} w_\theta(i)\,Q_i,$7, define

$P_\theta=\sum_{i\in E} w_\theta(i)\,Q_i,$8

and the random finite measure

$P_\theta=\sum_{i\in E} w_\theta(i)\,Q_i,$9

The theorem proves

$K$^^^^2query2^^^^

that is, equality in distribution at the prior level (&&&^^^^2query2^^^^&&&).

The representation is exact rather than asymptotic. It preserves the predictive distributions, the induced partition law, and the EPPF of the original infinite random measure. The proof is based on a telescoping identity and Tonelli’s theorem; in the summary this identity is written as

$K$^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^

The support of $K$2 is $K$3, with pmf $K$4. Its law depends jointly on the schedule $K$5 and the SSP weights $K$6, and arises from a slice-like decomposition over intervals $K$7 (&&&^^^^2query2^^^^&&&).

For proper SSPs the finite representation remains purely atomic: no residual $K$8 term appears in $K$9. If an SSP has $P_\theta=\sum_i w_\theta(i)Q_i$^^^^2query2^^^^, the theorem applies to its proper atomic part, while the diffuse residual remains unaffected. The representation is therefore a prior-level identity that converts an infinite paintbox into a mixture over finite paintboxes of random size $P_\theta=\sum_i w_\theta(i)Q_i$^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^, without any ad hoc truncation (&&&^^^^2query2^^^^&&&).

4. Canonical instances: Dirichlet, Pitman–Yor, and geometric stick-breaking

For the Dirichlet process, described in the synthesis as Poisson–Dirichlet $P_\theta=\sum_i w_\theta(i)Q_i$2 with $P_\theta=\sum_i w_\theta(i)Q_i$3, the EPPF is

$P_\theta=\sum_i w_\theta(i)Q_i$4

where $P_\theta=\sum_i w_\theta(i)Q_i$5 is the rising factorial. For $P_\theta=\sum_i w_\theta(i)Q_i$6, with $P_\theta=\sum_i w_\theta(i)Q_i$7 and $P_\theta=\sum_i w_\theta(i)Q_i$8,

$P_\theta=\sum_i w_\theta(i)Q_i$9

where $\mathcal K_l$^^^^2query2^^^^ is the generalized rising factorial (&&&^^^^2query2^^^^&&&).

Both priors admit stick-breaking weights

$\mathcal K_l$^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^

with

$\mathcal K_l$2

Under the stick-breaking schedule

$\mathcal K_l$3

the finite representation becomes

$\mathcal K_l$4

and

$\mathcal K_l$5

Marginalizing over $\mathcal K_l$6 recovers the original DP or PY prior, and the EPPF is preserved because the mixture identity reconstructs the law of $\mathcal K_l$7 (&&&^^^^2query2^^^^&&&).

A special case is geometric stick-breaking (GSB). If $\mathcal K_l$8, then

$\mathcal K_l$9

and

$(X,\mathcal B_X)$^^^^2query2^^^^

If $(X,\mathcal B_X)$^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^, the unconditional pmf is

$(X,\mathcal B_X)$2

This example makes explicit how an infinite decreasing-weight construction can induce a finite mixture with equal weights conditional on $(X,\mathcal B_X)$3 (&&&^^^^2query2^^^^&&&).

5. Posterior computation and empirical behavior

In mixture modeling, with kernel $(X,\mathcal B_X)$4 and observations $(X,\mathcal B_X)$5 i.i.d. from $(X,\mathcal B_X)$6, the finite representation supports standard finite-mixture machinery via allocation variables $(X,\mathcal B_X)$7 and per-observation truncations $(X,\mathcal B_X)$8. The generic hierarchical augmentation is:

• $(X,\mathcal B_X)$9; $$G(A)=\sum_{j=1}^\infty w_j\delta_{\theta_j}(A)+w_0G_0(A),\qquad A\in\mathcal B_X,$$^^^^2query2^^^^ i.i.d.
• $$G(A)=\sum_{j=1}^\infty w_j\delta_{\theta_j}(A)+w_0G_0(A),\qquad A\in\mathcal B_X,$$^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^ with $$G(A)=\sum_{j=1}^\infty w_j\delta_{\theta_j}(A)+w_0G_0(A),\qquad A\in\mathcal B_X,$$2.
• $$G(A)=\sum_{j=1}^\infty w_j\delta_{\theta_j}(A)+w_0G_0(A),\qquad A\in\mathcal B_X,$$3 with $$G(A)=\sum_{j=1}^\infty w_j\delta_{\theta_j}(A)+w_0G_0(A),\qquad A\in\mathcal B_X,$$4.
• $$G(A)=\sum_{j=1}^\infty w_j\delta_{\theta_j}(A)+w_0G_0(A),\qquad A\in\mathcal B_X,$$5 (&&&^^^^2query2^^^^&&&).

The corresponding joint kernel factorizes as

$$G(A)=\sum_{j=1}^\infty w_j\delta_{\theta_j}(A)+w_0G_0(A),\qquad A\in\mathcal B_X,$$6

with $$G(A)=\sum_{j=1}^\infty w_j\delta_{\theta_j}(A)+w_0G_0(A),\qquad A\in\mathcal B_X,$$7 and $$G(A)=\sum_{j=1}^\infty w_j\delta_{\theta_j}(A)+w_0G_0(A),\qquad A\in\mathcal B_X,$$8. This yields straightforward Gibbs updates. In conjugate cases,

$$G(A)=\sum_{j=1}^\infty w_j\delta_{\theta_j}(A)+w_0G_0(A),\qquad A\in\mathcal B_X,$$9

For stick-breaking variables $\theta_j$^^^^2query2^^^^, if $\theta_j$^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^ and $\theta_j$2, then

$\theta_j$3

where $\theta_j$4 and $\theta_j$5. For the DP this becomes

$\theta_j$6

If $\theta_j$7 is deterministic and independent of the stick-breaking variables, then

$\theta_j$8

where $\theta_j$9. For GSB,

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2query2^^^^

Allocations satisfy

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2id:(Mena et al., 30 Dec 2025) OR \2^^^^

and in the stick-breaking-dependent case PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2query22^^^^, so

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^3

For GSB, PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^4, hence PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^5 over PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^6 (&&&^^^^2query2^^^^&&&).

The PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^7-updates are especially transparent. In the schedule-dependent case,

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^8

If PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^9, then

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^

If PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2^^^^, then

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \22^^^^

If PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2^^^^3, then PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2^^^^4 with PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2^^^^5. For GSB, PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2^^^^6 with PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2^^^^7 (&&&^^^^2query2^^^^&&&).

Empirically, the finite samplers were compared against slice samplers on a simulated four-component Normal mixture and the galaxy data. In the simulated example PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2^^^^8, the true weights were PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2id:(Mena et al., 30 Dec 2025) OR \2^^^^9, with means PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \22query2^^^^ and sds PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \22id:(Mena et al., 30 Dec 2025) OR \2^^^^. The methods were DPFinite and DPSlice, and GSBFinite and GSBSlice. For moderate/large PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \222^^^^ and “natural” PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \22^^^^3, DPFinite and DPSlice recovered the density well and PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \22^^^^4 aligned closely with PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \22^^^^5; GSB-based methods slightly over-clustered. Runtime for PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \22^^^^6 iterations, evaluated over PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \22^^^^7 grid points, was: for PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \22^^^^8, DPFinite natural PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \22^^^^9s, PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2max_results2query2^^^^ PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2max_results2id:(Mena et al., 30 Dec 2025) OR \2^^^^s, DPSlice PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^32 PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^33s; GSBFinite natural PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^34s, PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^35 PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^36s, GSBSlice PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^37 PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^38s. For PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^39, DPFinite natural PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2sort_by2query2^^^^s, PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2sort_by2id:(Mena et al., 30 Dec 2025) OR \2^^^^ PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^42s, DPSlice PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^43s; GSBFinite natural PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^44s, PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^45 PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^46s, GSBSlice PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^47s. On the galaxy data PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^48, all methods captured multimodality; DPFinite and DPSlice stabilized around PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^49 clusters; small PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2submittedDate2query2^^^^ inflated PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2submittedDate2id:(Mena et al., 30 Dec 2025) OR \2^^^^ slightly (&&&^^^^2query2^^^^&&&).

6. Extensions, assumptions, and numerical considerations

Because the finite representation is a prior-level identity, it extends pointwise to non-exchangeable or covariate-dependent settings. If PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^52 is proper for each PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^53, with PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^54 almost surely, and if there exists a strictly decreasing PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^55, then one can introduce PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^56 with

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^57

and reweight

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^58

yielding PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^59. Algorithmically, this gives time- or covariate-indexed finite mixtures with exact preservation of the original non-exchangeable prior (&&&^^^^2query2^^^^&&&).

The stated assumptions are: a proper SSP, a diffuse base measure for atoms, and independence PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2sort_order2query2^^^^. The schedule PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2sort_order2id:(Mena et al., 30 Dec 2025) OR \2^^^^ may be deterministic or random, but it must be strictly decreasing to PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^62. The scope includes all proper SSPs, including Gibbs-type priors, normalized CRMs admitting stick-breaking forms, and stick-breaking processes with dependent lengths such as DSBw (&&&^^^^2query2^^^^&&&).

The main numerical guidance concerns the choice of schedule and the management of tail probabilities. Deterministic schedules such as PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^63 or PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^64 yield simple, stable PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^65-updates by closed-form inversion. Very small PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^66 can inflate PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^67, since more components become available. In the stick-breaking-dependent case, synchronization of PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^68-updates and tail mass PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^69 is important, and stable accumulators for PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2descending2query2^^^^ and tail sums are recommended to avoid floating-point drift (&&&^^^^2query2^^^^&&&).

A further scope qualification concerns interpretation of cluster counts. The occupied-cluster summary PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2descending2id:(Mena et al., 30 Dec 2025) OR \2^^^^ is data-driven and should not be interpreted as a direct estimate of a “true” finite PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^72. The exact finite representation yields exact inference for SSP mixtures; when the modeling objective is consistent recovery of a finite PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^73, the supplied synthesis states that mixtures of finite mixtures with PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^74 and explicit priors on PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^75 are appropriate (&&&^^^^2query2^^^^&&&).

7. Related and non-equivalent uses

Outside Bayesian nonparametrics, the same label has been attached to different kinds of mixture constructions. In multimodal learning, one such usage refers to a shared Gaussian mixture latent space for images and text. There the image-induced mixture is

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^76

while the text-induced mixture is

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^77

Textual compositionality is represented by mixing token-level Gaussians, and the training objective is

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^78

with

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^79

This model uses STN-based patch extraction, diagonal covariances in practice, and PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2arxiv_id2query2^^^^ in all experiments. On MultiCIFAR^^^^2id:(Mena et al., 30 Dec 2025) OR \2query2^^^^, the reported detection mAP values were WSDDN PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2arxiv_id2id:(Mena et al., 30 Dec 2025) OR \2^^^^ and CoMix PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^82; on MultiMNIST, WSDDN PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^83 and CoMix PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^84 (Alaniz et al., 2022).

A different usage arises in the analysis of shifted Gaussians through the King function. For dimensionless radial coordinate PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^85 and shift parameter PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^86, the King kernel is

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^87

where PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^88 is the modified spherical Bessel function. The paper establishes the King differential equation, a Gaussian-weighted self-adjoint operator, a unitary equivalence to the free radial Schrödinger operator, and a spectral representation on the imaginary branch. On the real branch,

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^89

is not orthogonal in

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2(Mena et al., 30 Dec 2025)2query2^^^^

but its linear span is dense. This yields discrete approximations of the form

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2(Mena et al., 30 Dec 2025)2id:(Mena et al., 30 Dec 2025) OR \2^^^^

with least-squares fitting in PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^92 (Wang et al., 5 Jun 2026).

Against these usages, the general theory of discrete mixture representations provides the broadest abstraction:

PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^93

That theory includes classical noncentral PRESERVED_PLACEHOLDER_^^^^2id:(Mena et al., 30 Dec 2025) OR \2^^^^94 mixtures, barycentric convexity, minimality and non-minimality phenomena, information inequalities, Rao–Blackwellization, and NPMLE for mixing probabilities, but explicitly notes that “King Mixture Representations” is not a standard term there (Baringhaus et al., 2022).

Taken together, these works indicate that the phrase does not name a single unified theory across fields. Its most specific and technically developed meaning is the exact finite-mixture reparametrization of proper species sampling processes compatible with Kingman’s paintbox, while other occurrences denote either mixture-based compositional embeddings or approximation schemes built from King functions.