Pith. sign in

REVIEW 3 major objections 4 minor 26 references

This paper proves that, in a linear squared-loss setting, explicit parameter adaptation and implicit routing (in-context learning, retrieval, mixture-of-experts) both reduce to kernel ridge regression on joint input-context features.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-07-31 23:48 UTC pith:V6SQJFAG

load-bearing objection Useful survey that overclaims the precision of its central equivalence; the expository kernel bridge is sound in the linear trained-head sense, but the abstract's 'prove equivalent' outstrips what Proposition 1(B) actually shows. the 3 major comments →

arxiv 2607.23304 v1 pith:V6SQJFAG submitted 2026-07-25 stat.ML cs.LGstat.ME

Context-Adaptive Inference: A Unified Statistical and Foundation-Model View

Yue Yao , Caleb N. Ellington , Jingyun Jia , Baiheng Chen , Dong Liu , Rikhil Rao , Jiaqi Wang , Samuel Wales-McGrath
show 4 more authors
Yixin Yang Zhiyuan Li Eric P. Xing Ben Lengerich
This is my paper
classification stat.ML cs.LGstat.ME MSC 62G0862J0768T07
keywords context-adaptive inferencevarying-coefficient modelsin-context learningkernel ridge regressionmixture-of-expertsmeta-learningretrieval-augmented generationunification
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper argues that the many ways models adapt to context — classical varying-coefficient regression, meta-learned fast adaptation, retrieval-augmented prompting, and mixture-of-experts routing — are not separate tricks but variants of a single estimator: kernel ridge regression on features that pair the input with the context. Under squared loss with linear prediction heads and fixed features, it proves that explicit parameter adaptation and implicit attention-based routing coincide with the same kernel estimator. The paper then turns that equivalence into a practical toolbox: design choices such as which context encoder to use, what regularizer to impose, and how much compute to spend become shared tuning knobs across all three traditions. A sympathetic reader would care because it means insights, failure modes, and fixes from one approach transfer to the others.

Core claim

The central claim is that explicit and implicit context adaptation are the same learning rule. Concretely, the paper's Proposition 1 shows that (A) a varying-coefficient model fit by locally weighted ridge regression and (B) a single linear attention layer performing in-context learning both reduce to kernel ridge regression on the joint feature vector formed by the Kronecker product of the input and a context encoding. In the explicit case the kernel is a user-chosen similarity on contexts; in the implicit case it is a learned dot-product kernel (or, for trained networks, the network's NTK). Retrieval, gating, and weighting — the mechanisms behind RAG and mixture-of-experts — are corollarie

What carries the argument

The engine of the argument is the context encoder φ(c) and the joint feature map x ⊗ φ(c), combined with kernel ridge regression. Proposition 1 uses ridge/dual duality to show that any weighted ridge solution with context-dependent support equals KRR on these joint features (part A), and that a linear attention layer with linear query/key/value maps and a linear readout induces exactly the same form, with a learned dot-product kernel on the same joint features (part B). The identity does the unificatory work: it turns 'does this model adapt?' into 'which kernel and which empirical measure is the model using?' — making context encoder, regularizer, and compute the three shared controls.

Load-bearing premise

The bridge relies on the linear, squared-loss regime: linear prediction heads, fixed features, and either a single linear attention layer or a network in its NTK-linearized regime; if real in-context learning in foundation models is not well approximated by that regime, the equivalence is a statement about idealized models rather than deployed systems.

What would settle it

Train a transformer to perform in-context learning on a nonlinear task (e.g., logistic regression or data with a quadratic head) and test whether its predictions can be reproduced by any fixed kernel ridge regression on the joint features. If no single kernel matches the model across multiple distinct support sets, the fixed-kernel claim fails outside the linear regime; conversely, finding such a kernel would support the unified view.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • If true, a single estimation theory covers varying-coefficient models, meta-learned adaptation, and in-context learning; results about kernel regression (bias-variance, regularization, consistency) transfer to transformer-based ICL and vice versa.
  • Design choices in one family become interpretable in the other: a prompt's support set plays the role of a kernel neighborhood, and a retriever or router plays the role of a kernel/measure choice.
  • The equivalence gives an audit path: for linearized regimes, one can ask what kernel an in-context model has learned and check whether it matches the intended similarity structure.
  • The proposed evaluation metrics (adaptation-efficiency, routing stability, context-specific robustness) become meaningful across all three traditions, since they measure properties of the shared estimator.
  • The paper's 'context → parameters → prediction' objective provides a common language for comparing methods that were previously incommensurable.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Extension: If the equivalence holds beyond the toy regime, then scaling laws for context-adaptive models — the paper's 'horizontal vs. vertical scaling' — could be derived from classical kernel-regression learning curves, letting practitioners predict when adding new contexts beats adding more data per context.
  • The paper leaves implicit a direct empirical check: train a transformer on linear regression tasks, extract its effective attention kernel, and test whether it equals a fixed dot-product kernel on joint features; a positive result would make the equivalence experimentally testable in deployed models, while a negative result would show exactly where the linear regime ends.
  • Extension: In the nonlinear regime the paper says the fixed-kernel view breaks; one testable hypothesis is that it is replaced by an input-dependent or prompt-dependent kernel (e.g., a kernel whose weights renormalize with representations), which would preserve the unified view in a generalized form.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper proposes a unified framework for 'context-adaptive inference,' in which a context c is mapped to adapted parameters θ(c) and prediction proceeds via f(x; θ(c)). It surveys three traditions—explicit statistical adaptation (varying-coefficient models, mixed models, contextualized models), meta-learning/transfer, and implicit adaptation in foundation models (attention, prompting, retrieval, mixture-of-experts). The central formal claim is Proposition 1: under squared loss, linear prediction heads, and fixed features, both explicit varying-coefficient models and implicit linear attention/in-context learning are equivalent to kernel ridge regression (KRR) on joint features of inputs and context z = φ(c) ⊗ x. The paper then distills design principles, proposes evaluation metrics (adaptation efficiency, routing stability, context-specific robustness), and discusses applications and open problems.

Significance. If the equivalence in Proposition 1 were rigorously established, the paper would provide a genuinely useful bridge between classical statistics and foundation-model research, and the proposed evaluation metrics could guide practical deployment of adaptive systems. The survey portions—especially the taxonomy of explicit adaptivity and the failure-mode analysis—are valuable and well-grounded in the literature. The explicit half of the bridge (part (A)) is standard and correct. However, the implicit half (part (B)) as stated is not supported: a single fixed-weight attention forward pass cannot in general equal the KRR estimator, because attention computes a data-dependent weighted average without the inverse Gram matrix. The proof overview itself shifts to trained-head ridge and NTK linearization, which are different mechanisms from fixed-weight in-context inference. Since the abstract's 'mathematically equivalent' claim rests on this part, the central theorem needs substantive correction.

major comments (3)
  1. [Section 5, Proposition 1(B)] The claim that a single linear attention layer with fixed linear query/key/value maps and a linear readout 'equals KRR with kernel K' is not justified and, as stated, is false. A softmax or linear attention forward pass computes a weighted average of value vectors, with weights determined by query–key similarities. KRR, in contrast, requires solving (K+λI)⁻¹Y. For fixed W_Q, W_K, W_V, there is no mechanism to encode the support-set-dependent inverse (K+λI)⁻¹. The proof overview does not repair this: it appeals to (i) 'fixed attention + trained linear head = ridge on fixed features' and (ii) NTK linearization. Both are different from a single fixed-weight forward pass. Please provide a precise statement with any auxiliary assumptions (e.g., trained head, iterative updates, special constructions of W_V), and reconcile the statement with the abstract's 'mathematically equivalent' wording.
  2. [Section 5, Proof Overview vs. Section 4 definition of ICL] There is an internal inconsistency in what 'implicit' means. Section 4 defines in-context learning as adaptation 'entirely within a single forward pass, without any updates to the model's weights.' Yet the proof overview for Proposition 1(B) relies on 'fixed attention + trained linear head' and 'attention parameters trained in the linearized/NTK regime.' Training a head or fine-tuning attention parameters on the support set is explicit parameter adaptation, not fixed-weight in-context inference. This conflation is load-bearing because the paper's advertised bridge depends on it. Please disambiguate the two settings and restrict the equivalence claim to the setting actually proved.
  3. [Appendix A (referenced in Section 5)] The text repeatedly states 'Full proof in Appendix A,' but the version under review does not include Appendix A. The proof sketch in the main text is not sufficient to establish Proposition 1(B): it does not define the kernel K, does not specify how softmax normalization is handled, and does not show how a fixed-parameter attention layer can reproduce the KRR inverse. Since the central claim of the paper depends on this proof, the appendix must be supplied and its contents must be consistent with the main-text statement.
minor comments (4)
  1. [Throughout] Many inline equations and symbols are missing from the rendered text (e.g., Eq. (1) in Section 5, the adaptation-efficiency and bias-variance formulas in Section 7, and parts of Proposition 1). This makes verification unnecessarily difficult. Please ensure the final version has complete, correctly typeset mathematics.
  2. [Section 4] The discussion of implicit adaptivity contains repeated sentences nearly verbatim (e.g., 'This phenomenon is known as implicit adaptivity' appears multiple times). Tighten the prose without changing the content.
  3. [Table 1] The survey comparison table would be more useful if it explicitly listed the present paper's contribution in the 'Gap Relative to This Work' column, rather than only describing prior surveys.
  4. [Section 7] The formal metrics (adaptation efficiency, routing stability, context-specific robustness) are introduced with formulas that are not fully legible in the current version. Consider adding a plain-language interpretation of each metric and a small illustrative example.

Circularity Check

0 steps flagged

No significant circularity: the central bridge is explicitly expository and rests on external prior work; self-citations frame the review but do not carry the proof.

full rationale

The paper's load-bearing claim is Proposition 1: explicit varying-coefficient models and linear in-context learning coincide with kernel ridge regression on joint features under squared loss, linear heads, and fixed features. The manuscript itself disclaims novelty for this proposition: "Proposition 1 is expository: part (A) is standard ridge–kernel duality on joint features; part (B) follows from (i) fixed attention + trained linear head = ridge on fixed features and (ii) NTK linearization = kernel regression with the network's NTK." The implicit half is attributed to external, independently checkable prior work (Akyürek et al. 2022; von Oswald et al. 2023; Garg et al. 2023; Dai et al. 2023; Reuter et al. 2025 and NTK analyses), not to the present authors' own fitted results or definitions. The authors also state "See transformer ICL as classical estimators [136,137,138] and NTK analyses [139,140]," further indicating that the equivalence is imported from the literature rather than derived from a self-citation chain. The paper's heavy use of self-citations (contextualized models, personalized regression, NOTMAD) is for framing, examples, and tooling; these works are not used to prove the equivalence or to define the kernel in a way that presupposes the conclusion. No fitted parameter is renamed as a prediction, no uniqueness theorem from the authors is invoked to force a choice, and no ansatz is smuggled in through self-citation. The detailed proof is deferred to Appendix A, but the proof overview and the explicit reliance on external results mean this is an omitted-proof concern, not circularity. The remaining concerns about whether a softmax-attention forward pass equals KRR with a ridge inverse are correctness/scope questions, not circularity: they do not show that the paper's conclusion is equivalent to its inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 6 axioms · 0 invented entities

The equivalence theorem depends on linearity, exchangeability, and NTK linearization; the survey's broader principles are not derived from these axioms alone and are supported mainly by citations. No numeric free parameters are fitted, and the paper introduces no new physical or mathematical entities beyond organizing concepts such as θ(c).

axioms (6)
  • domain assumption Prediction model is linear in fixed features with squared loss: f(x;θ)=θ^T φ(x).
    Proposition 1's equivalence rests on this restriction; the paper's own Limitations section says non-quadratic losses and nonlinear heads break the fixed-kernel view.
  • domain assumption Within-context samples are exchangeable/i.i.d. given context c.
    Stated in Section 5 Proof Overview item 1; needed to treat the weighted support set as a valid empirical distribution.
  • standard math Regularity of coefficient functions and bounded, locally-normalized retrieval weights.
    Section 5 Proof Overview item 2; technical condition for the weighted kernel ridge equivalence.
  • standard math Risk is convex in model outputs and the minimizer is unique or stable.
    Section 5 Proof Overview item 3; needed for identifiability and stability of the estimator.
  • ad hoc to paper Implicit in-context learning is represented by a single linear attention layer with linear readout, and trained attention is analyzed in the NTK linearization regime.
    Section 5, Prop 1(B) and the NTK discussion; a strong simplification because real transformers use nonlinear attention, softmax weights, and MLP blocks.
  • domain assumption For causal interpretation of heterogeneity, there is no unmeasured confounding given context.
    Sections 1 and 6 state that recovering subgroup relationships requires conditioning on the contextual side information; without this, Simpson's-paradox-type reversals cannot be resolved.

pith-pipeline@v1.3.0-alltime-deepseek · 42129 in / 13463 out tokens · 135135 ms · 2026-07-31T23:48:07.980886+00:00 · methodology

0 comments
read the original abstract

Modern predictive systems are expected to adapt their behavior to the specific situation they are facing. A clinical model should not treat every patient the same; a retrieval-augmented model should change its answer when given different evidence; a mixture-of-experts model should route different inputs to different experts. We call this capability context-adaptive inference: before predicting, the system uses information about the current context to specialize its parameters or computation for that instance. This article provides a unified view of context-adaptive inference across three traditions that are usually treated separately: (i) explicit adaptation in statistics (e.g. varying-coefficient models, local regression, hierarchical sharing), (ii) rapid task-specific adaptation in meta-learning and transfer, and (iii) implicit adaptation in large foundation models via prompting, retrieval, and expert routing. We formalize these approaches under a common objective: to map context $c$ to adapted parameters $\theta(c)$, then to predict via $f(x; \theta(c))$. Under squared loss, linear prediction heads, and fixed features, we prove that explicit parameter adaptation and implicit routing are mathematically equivalent to kernel ridge regression on joint features of inputs and context. Building on this bridge, we propose practical design principles and evaluation metrics including adaptation-efficiency, routing stability, and context-specific robustness to guide when to specialize, how to constrain that specialization, and how to audit context-adaptive models in deployment. Finally, we identify open problems in identifiability, robustness under distribution shift, and efficient large-scale adaptation, outlining design principles for methods that are scalable, reliable, and transparent in real-world settings.

Figures

Figures reproduced from arXiv: 2607.23304 by Baiheng Chen, Ben Lengerich, Caleb N. Ellington, Dong Liu, Eric P. Xing, Jiaqi Wang, Jingyun Jia, Rikhil Rao, Samuel Wales-McGrath, Yixin Yang, Yue Yao, Zhiyuan Li.

Figure 1
Figure 1. Figure 1: Failure Modes of Population Models. Illustrative schematics of common failure types when fitting a single global model to heterogeneous data. (A) Mode Collapse: the dominant group drives the fit, underrepresenting the minority. (B) Outlier Sensitivity: extreme points distort the global line, shifting predictions away from the majority. (C) Phantom Populations: the global fit represents no actual subgroup, … view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the theoretical bridge. Three traditions—Statistics (varying-coefficients, local smoothing, hierarchical sharing), Meta-learning (bilevel training, fast adaptation, hypernetworks), and Foundation models (prompted inference / in-context learning)—feed into a unified context→parameters view. The bridge formalizes this connection and highlights shared tuning knobs: context information, inductive b… view at source ↗
Figure 3
Figure 3. Figure 3: Multi-task learning as explicit adaptivity. In traditional MTL (left), each task has its own head on top of shared layers. In context-flagged models (right), the task identity is provided as an input, enabling a shared model to adapt across tasks. The canonical formalism for explicit adaptivity is the varying-coefficient model (VCM), which writes each regression coefficient as a function of context [3,63]:… view at source ↗
Figure 4
Figure 4. Figure 4: A spectrum of context awareness in modeling, showing global, grouped, smooth, and latent models Each step forward increases the power of context-specific, personalized inference by borrowing strength from related samples and groups. Each step also lowers the amount of data that must be collected for any single context, because information flows in from neighboring contexts rather than being estimated in is… view at source ↗
Figure 5
Figure 5. Figure 5: Hierarchical splits of context enable multi-level adaptivity. Explicit adaptivity can partition the context space into piecewise models, with parameters indexed both by context and task identity . Such splits allow sample-level heterogeneity to be captured within contexts, while high-level partitions mimic task boundaries and enable task switching. Dissolving the Boundaries with Kernels Partitions, however… view at source ↗
Figure 6
Figure 6. Figure 6: Compositional inference: nonparametric flexibility from parametric context-specific models. (A) Overall conditional . (B) Context-specific components for latent subgroups . (C) Recombination via marginalization . (D) Aggregated distribution showing how structured parametric pieces yield multimodal, nonparametric-like behavior [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Patterns of missingness as context. Each dataset (e.g., cohort with labs, cohort with vitals, cohort with imaging) provides a different subset of measurements. Context-adaptive models allow integration by conditioning on measurement availability, enabling learning from fewer samples with more heterogeneous features. Key Theoretical Advances The guarantees available for these methods reflect the assumptions… view at source ↗
Figure 8
Figure 8. Figure 8: Evolution of Statistical Estimation. Classical estimators (left) require sufficiently large context-specific datasets for single-instance learning. Meta-learning and transfer learning (middle) decrease the need for context-specific data collection, but still perform a few rounds of gradient descent on context-specific data. Finally, amortized estimators (right) only use a representation of the data context… view at source ↗
Figure 9
Figure 9. Figure 9: Heterogeneous treatment effects. Left: average treatment effect (ATE) conditional on , implicitly assuming homogeneity across contexts. Right: conditional average treatment effect (CATE) that allows treatment effects to vary systematically with additional context . In Eq. 1, this principle is encoded by : a signal of heterogeneity is what makes the support set localized rather than an arbitrary neighborhoo… view at source ↗
Figure 10
Figure 10. Figure 10: Context-adaptive models use context signals to understand specific subpopulations or tasks (e.g. modeling thyroid cancer patients) within broader, heterogeneous populations or task sets (e.g. all cancer patients) during training. At inference time, context localizes the model to a specific predictive task. Context-adaptive models have higher [PITH_FULL_IMAGE:figures/full_fig_p032_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Failure Modes of Context-Adaptive Models. (A) Spurious Adaptation: the fit tracks a feature that predicts in the training environment but reverses under shift. (B) Overfitting in Low-Data Contexts: adaptation follows noise rather than signal. (C) Modularity Mis-Specification: incorrect partitions obscure the true structure. (D) Feedback Loops: adaptive decisions reshape the very data they rely on. Implica… view at source ↗
Figure 12
Figure 12. Figure 12: From Implicit to Explicit Adaptivity. A black-box model (left) represents implicit adaptation, which is hidden and opaque. Making adaptivity explicit (right) exposes structured components that can be inspected and controlled. The axes below highlight the trade-offs between fidelity and interpretability, local and global scope, and approximation and control. Approaches Efforts to make implicit adaptation e… view at source ↗
Figure 13
Figure 13. Figure 13: Contextualized networks enable inference of archetypal and sample-specific mixtures, unlocking new biological insights such as mechanisms of disease, disease heterogeneity, structural hazards, and driver mutations [PITH_FULL_IMAGE:figures/full_fig_p048_13.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

26 extracted references · 6 canonical work pages · 6 internal anchors

  1. [1]

    Local Likelihood Estimation Robert Tibshirani, Trevor Hastie Journal of the American Statistical Association (1987-06) https://doi.org/g9krnh DOI: 10.1080/01621459.1987.10478466 Estimation of Genetic Parameters Charles R Henderson Annals of Mathematical Statistics (1950) Varying-Coefficient Models Trevor Hastie, Robert Tibshirani Journal of the Royal Stat...

    arXiv 1987

  2. [21]

    DOI: 10.48550/arxiv.2307.13721 A Comprehensive Survey on Pretrained Foundation Models: A History from BERT to ChatGPT Ce Zhou, Qian Li, Chen Li, Jun Yu, Yixin Liu, Guangjing Wang, Kai Zhang, Cheng Ji, Qiben Yan, Lifang He, … Lichao Sun arXiv (2023) https://doi.org/g8vjrk DOI: 10.48550/arxiv.2302.09419 The Design of Experiments Ronald A Fisher Oliver & Boy...

    Pith/arXiv arXiv doi:10.48550/arxiv.2307.13721 2023

  3. [31]

    DOI: 10.1037/h0040957 · PMID: 13245896 A Method of Estimating Comparative Rates from Clinical Data. Applications to Cancer of the Lung, Breast, and Cervix JNCI: Journal of the National Cancer Institute (1951-06) https://doi.org/g96wsb DOI: 10.1093/jnci/11.6.1269 Recovery of inter-block information when block sizes are unequal HD PATTERSON, R THOMPSON Biom...

    arXiv 1951

  4. [42]

    Representation Learning: A Review and New Perspectives Y Bengio, A Courville, P Vincent IEEE Transactions on Pattern Analysis and Machine Intelligence (2013-08) https://doi.org/ f42hw4 DOI: 10.1109/tpami.2013.50 · PMID: 23787338 Multitask Learning Rich Caruana Machine Learning (1997-07) https://doi.org/d3gsgj DOI: 10.1023/a:1007379606734 A Survey on Trans...

    Pith/arXiv arXiv doi:10.48550/arxiv.1606.04080 2013

  5. [53]

    DOI: 10.1007/978-3-540-28645-5_29 Early Drift Detection Method Manuel Baena-García, José del Campo-Ávila, Raul Fidalgo, Albert Bifet, Ricard Gavalda, Rafael Morales-Bueno Fourth International Workshop on Knowledge Discovery from Data Streams (2006) New ensemble methods for evolving data streams Albert Bifet, Geoff Holmes, Bernhard Pfahringer, Richard Kirk...

    arXiv 2006

  6. [62]

    Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A Rusu, Joel Veness, Marc G Bellemare, Alex Graves, Martin Riedmiller, Andreas K Fidjeland, Georg Ostrovski, … Demis Hassabis Nature (2015-02-25) https://doi.org/gc3h75 DOI: 10.1038/nature14236 · PMID: 25719670 Compression, restoration, resampling, ‘compressive sensing’: fast transforms in digital im...

    arXiv 2015

  7. [72]

    arXiv (2022) https://doi.org/gthszs DOI: 10.48550/arxiv.2203.15556 Publication Trends on the Varying Coefficients Model: Estimating the Actual (Under)Utilization of a Highly Acclaimed Method for Studying Statistical Interactions Assaf Botzer Publications (2025-04-07) https://doi.org/g9t2rq DOI: 10.3390/publications13020019 Estimating time-varying networks...

    Pith/arXiv arXiv doi:10.48550/arxiv.2203.15556 2022

  8. [83]

    Bayesian Inference of Multiple Gaussian Graphical Models Christine Peterson, Francesco C Stingo, Marina Vannucci Journal of the American Statistical Association (2015-01-02) https://doi.org/f69dnj DOI: 10.1080/01621459.2014.896806 · PMID: 26078481 · PMCID: PMC4465207 Learning Sample-Specific Models with Low-Rank Personalized Regression Benjamin Lengerich,...

    arXiv 2015

  9. [93]

    arXiv (2024) https://doi.org/g9582j DOI: 10.48550/arxiv.2410.07229 Spatially Varying Coefficient Models for Estimating Heterogeneous Mixture Effects Jacob Englert, Howard Chang arXiv (2025) https://doi.org/g9582q DOI: 10.48550/arxiv.2502.14651 Network Varying Coefficient Model Xinyan Fan, Kuangnan Fang, Wei Lan, Chih-Ling Tsai Journal of the American Stat...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2410.07229 2024

  10. [103]

    HyperNetworks David Ha, Andrew Dai, Quoc V Le arXiv (2016) https://doi.org/hb8768 DOI: 10.48550/arxiv.1609.09106 FiLM: Visual Reasoning with a General Conditioning Layer Ethan Perez, Florian Strub, Harm de Vries, Vincent Dumoulin, Aaron Courville arXiv (2017) https://doi.org/hb8769 DOI: 10.48550/arxiv.1709.07871 Contextualized Machine Learning Benjamin Le...

    Pith/arXiv arXiv doi:10.48550/arxiv.1609.09106 2016

  11. [112]

    Ram Dyuthi Sristi, Ofir Lindenbaum, Shira Lifshitz, Maria Lavzin, Jackie Schiller, Gal Mishne, Hadas Benisty arXiv (2023) https://doi.org/gt68jh DOI: 10.48550/arxiv.2312.14254 Contextual Explanation Networks Maruan Al-Shedivat, Avinava Dubey, Eric P Xing arXiv (2017) https://doi.org/gt68h9 DOI: 10.48550/arxiv.1705.10301 Cell-Level Virtual Screening Caleb ...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2312.14254 2023

  12. [122]

    Auto-Encoding Variational Bayes Diederik P Kingma, Max Welling arXiv (2013) https://doi.org/gpp5xv DOI: 10.48550/arxiv.1312.6114 Meta-Learning in Neural Networks: A Survey Timothy Hospedales, Antreas Antoniou, Paul Micaelli, Amos Storkey arXiv (2020) https://doi.org/g958zx DOI: 10.48550/arxiv.2004.05439 Model-Agnostic Meta-Learning for Fast Adaptation of ...

    Pith/arXiv arXiv doi:10.48550/arxiv.1312.6114 2013

  13. [132]

    An Explanation of In-context Learning as Implicit Bayesian Inference Sang Michael Xie, Aditi Raghunathan, Percy Liang, Tengyu Ma arXiv (2021) https://doi.org/gtkkfs DOI: 10.48550/arxiv.2111.02080 In-Context Learning Strategies Emerge Rationally Daniel Wurgaft, Ekdeep Singh Lubana, Core Francisco Park, Hidenori Tanaka, Gautam Reddy, Noah D Goodman arXiv (2...

    Pith/arXiv arXiv doi:10.48550/arxiv.2111.02080 2021

  14. [143]

    DOI: 10.48550/arxiv.2212.07677 What Can Transformers Learn In-Context? A Case Study of Simple Function Classes Shivam Garg, Dimitris Tsipras, Percy Liang, Gregory Valiant arXiv (2022) https://doi.org/g9t22c DOI: 10.48550/arxiv.2208.01066 Can Transformers Learn Full Bayesian Inference in Context? Arik Reuter, Tim GJ Rudner, Vincent Fortuin, David Rügamer a...

    Pith/arXiv arXiv doi:10.48550/arxiv.2212.07677 2022

  15. [153]

    DOI: 10.48550/arxiv.2405.16075 LFME: A Simple Framework for Learning from Multiple Experts in Domain Generalization Liang Chen, Yong Zhang, Yibing Song, Zhiqiang Shen, Lingqiao Liu arXiv (2024) https://doi.org/g9582n DOI: 10.48550/arxiv.2410.17020 Scalable Multi-Domain Adaptation of Language Models using Modular Experts Peter Schafhalter, Shun Liao, Yanqi...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2405.16075 2024

  16. [163]

    Unsupervised Learning via Meta-Learning Kyle Hsu, Sergey Levine, Chelsea Finn arXiv (2018) https://doi.org/g958zs DOI: 10.48550/arxiv.1810.02334 Bayesian scaling laws for in-context learning Aryaman Arora, Dan Jurafsky, Christopher Potts, Noah D Goodman arXiv (2024) https://doi.org/g9582m DOI: 10.48550/arxiv.2410.16531 Shortcut Learning in Deep Neural Net...

    Pith/arXiv arXiv doi:10.48550/arxiv.1810.02334 2018

  17. [174]

    Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers Stephen Boyd, Neal Parikh, Eric Chu, Borja Peleato, Jonathan Eckstein Foundations and Trends® in Machine Learning (2011-07-26) https://doi.org/d3kztk DOI: 10.1561/2200000016 Domain Adaptation under Missingness Shift Helen Zhou, Sivaraman Balakrishnan, Zac...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2211.02093 2011

  18. [184]

    Why Should I Trust You?

    Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (2016-08-13) https://doi.org/gdp84q DOI: 10.1145/2939672.2939785 Towards A Rigorous Science of Interpretable Machine Learning Finale Doshi-Velez, Been Kim arXiv (2017) https://doi.org/h3cz DOI: 10.48550/arxiv.1702.08608 Rethinking Explainable Machine Learnin...

    arXiv 2016

  19. [194]

    beta-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework Irina Higgins, Loic Matthey, Arka Pal, Christopher Burgess, Xavier Glorot, Matthew Botvinick, Shakir Mohamed, Alexander Lerchner International Conference on Learning Representations (2017) https://openreview.net/forum? id=Sy2fzU9gl Deep Variational Information Bottleneck Ale...

    Pith/arXiv arXiv doi:10.48550/arxiv.1612.00410 2017

  20. [204]

    arXiv (2021) https://doi.org/gs8cqf DOI: 10.48550/arxiv.2104.08696 Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet Adly Templeton, Tom Conerly, Jonathan Marcus, Jack Lindsey, Trenton Bricken, Brian Chen, Adam Pearce, Craig Citro, Emmanuel Ameisen, Andy Jones, … Tom Henighan arXiv (2026) https://doi.org/hcbnz7 DOI: 10.48550/...

    Pith/arXiv arXiv doi:10.48550/arxiv.2104.08696 2021

  21. [216]

    Conditional Variance Penalties and Domain Shift Robustness Christina Heinze-Deml, Nicolai Meinshausen arXiv (2019-04-16) https://arxiv.org/abs/1710.11469 Causal inference using invariant prediction: identification and confidence intervals Jonas Peters, Peter Bühlmann, Nicolai Meinshausen arXiv (2024-04-26) https://arxiv.org/abs/1501.01332 Towards Deep Le...

    Pith/arXiv arXiv 2019

  22. [226]

    Statistics in Medicine (2022-10) https://doi.org/g96dmg DOI: 10.1002/sim.9582 · PMID: 36181392 · PMCID: PMC9931182 Dynamic effects of increasing heterogeneity in financial markets Ahmad K Naimzada, Giorgio Ricchiuti Chaos, Solitons & Fractals (2009-08) https://doi.org/bfbqxn DOI: 10.1016/j.chaos.2008.07.022 Bayesian Forecasting in Economics and Financ...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2212.03471 2022

  23. [237]

    arXiv (2026-06-10) https://arxiv.org/abs/2409.08354 Bayesian Models for Joint Selection of Features and Auto-Regressive Lags: Theory and Applications in Environmental and Financial Forecasting Alokesh Manna, Sujit K Ghosh arXiv (2025-08-18) https://arxiv.org/abs/2508.10055 Chaotic Bayesian Inference: Strange Attractors as Risk Models for Black Swan Events...

    Pith/arXiv arXiv 2026

  24. [247]

    LMPriors: Pre-Trained Language Models as Task-Specific Priors Kristy Choi, Chris Cundy, Sanjari Srivastava, Stefano Ermon arXiv (2022) https://doi.org/g9t22d DOI: 10.48550/arxiv.2210.12530 AdapterFusion: Non-Destructive Task Composition for Transfer Learning Jonas Pfeiffer, Aishwarya Kamath, Andreas Rücklé, Kyunghyun Cho, Iryna Gurevych arXiv (2020) https...

    Pith/arXiv arXiv doi:10.48550/arxiv.2210.12530 2022

  25. [257]

    TabPFN: A Transformer That Solves Small Tabular Classification Problems in a Second Noah Hollmann, Samuel Müller, Katharina Eggensperger, Frank Hutter arXiv (2022) https://doi.org/g9t22b DOI: 10.48550/arxiv.2207.01848 Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing Pengfei Liu, Weizhe Yuan, Jinlan Fu...

    Pith/arXiv arXiv doi:10.48550/arxiv.2207.01848 2022

  26. [258]

    DOI: 10.48550/arxiv.2406.19598 Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models Qizheng Zhang, Changran Hu, Shubhangi Upasani, Boyuan Ma, Fenglu Hong, Vamsidhar Kamanuru, Jay Rainton, Chen Wu, Mengmeng Ji, Hanchen Li, … Kunle Olukotun arXiv (2025) https://doi.org/g96dmr DOI: 10.48550/arxiv.2510.04618 Appendix A This append...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2406.19598 2025