Publications
13 papers: 7 published, 1 accepted, 5 preprints; 1 more in preparation, not counted. Each entry gives my position in the author list and, separately, what I contributed.
COMPACT Collaboration author lists are ordered first by contribution, then alphabetically. Long author lists are abbreviated; full lists are on arXiv, the journal page and INSPIRE-HEP.
Citation metrics
- Citations
- 122
- h-index
- 7
- i10-index
- 6
Google Scholar, as of 7 November 2026 · View profile
Published
-
Fourth author
The topology of the Universe
Nature Astronomy, Review Article (advance online publication) · Published online 11 Sep 2026
An invited review of what observations currently say about the global shape of the Universe, and how coming surveys could test it.
Reviews the observational status of cosmic topology and the prospects for CMB searches (LiteBIRD, Taurus) and large-scale-structure searches.
My contribution: Contributed to the CMB analysis, prepared the relevant plots, and helped develop the discussion of covariance analysis in cosmic-topology studies.
BibTeX for The topology of the Universe
% Abbreviated author list; see the DOI or INSPIRE-HEP for the full record. @article{Copi:2026review, author = {Copi, Craig J. and Mihaylov, Deyan P. and Negro, Anna and Samandar, Amirhossein and Starkman, Glenn D. and others}, title = {The topology of the {Universe}}, journal = {Nature Astronomy}, year = {2026}, doi = {10.1038/s41550-026-02930-6}, eprint = {2606.24886}, archivePrefix = {arXiv}, primaryClass = {astro-ph.CO} } -
Second author Lead author
Cosmic topology. Part IIb. Eigenmodes, correlation matrices, and detectability of non-orientable Euclidean manifolds
JCAP 07 (2026) 004 · 2026
Works out what the CMB would look like in each flat, finite universe that is non-orientable, and how detectable each would be.
Eigenmodes, non-diagonal aℓm covariance matrices, and cosmic-variance-limited detectability for the non-orientable Euclidean topologies E7–E10, E13–E15 and E17.
My contribution: Lead author: made a full first-author contribution to the research (second in the printed author list).
BibTeX for Cosmic topology IIb
% Abbreviated author list; see the DOI or INSPIRE-HEP for the full record. @article{Copi:2026iib, author = {Copi, Craig J. and Samandar, Amirhossein and Starkman, Glenn D. and {Carr{\'o}n Duque}, Javier and Akrami, Yashar and Anselmi, Stefano and Jaffe, Andrew H. and Kosowsky, Arthur and Cornet-Gomez, Fernando and Eskilt, Johannes R. and others}, title = {Cosmic topology. {Part IIb}. {Eigenmodes}, correlation matrices, and detectability of non-orientable {Euclidean} manifolds}, journal = {Journal of Cosmology and Astroparticle Physics}, volume = {2026}, number = {07}, pages = {004}, year = {2026}, doi = {10.1088/1475-7516/2026/07/004}, eprint = {2510.05030}, archivePrefix = {arXiv}, primaryClass = {astro-ph.CO} } -
Second author Lead author
Don't Pass@k: A Bayesian Framework for Large Language Model Evaluation
ICLR 2026, main conference · 2026
Replaces the Pass@k score used to compare language models with a Bayesian estimate that comes with error bars.
Outcomes are modeled as categorical with a Dirichlet prior, giving a closed-form posterior mean and uncertainty for any weighted rubric. Compared with Pass@k and avg@N, the estimate converges faster and gives more stable rankings at far smaller sample counts.
My contribution: As a lead author, co-developed the Bayesian framework for LLM evaluation, including posterior estimates and credible intervals for finite-sample uncertainty.
BibTeX for Don't Pass@k
@inproceedings{Hariri:2026passk, author = {Hariri, Mohsen and Samandar, Amirhossein and Hinczewski, Michael and Chaudhary, Vipin}, title = {Don't {Pass@k}: A {Bayesian} Framework for Large Language Model Evaluation}, booktitle = {Proceedings of the Fourteenth International Conference on Learning Representations (ICLR)}, year = {2026}, eprint = {2510.04265}, archivePrefix = {arXiv}, primaryClass = {cs.AI} } -
First author Lead author
Cosmic topology. Part IIIb. Eigenmodes and correlation matrices of spin-2 perturbations in orientable Euclidean manifolds
JCAP 08 (2025) 015 · 2025
Computes, for the first time, how primordial gravitational waves (tensor perturbations) would carry the imprint of a finite, flat universe into the CMB.
Derives spin-2 Laplacian eigenmodes and the corresponding CMB correlation matrices for orientable compact Euclidean manifolds. Over the multipole range computed, the topological information in tensor-induced anisotropies (measured by the KL divergence) keeps growing with ℓ, rather than saturating as quickly as in the scalar case.
My contribution: Led the analytical derivation of spin-2 eigenmodes and CMB covariance matrices, together with the numerical implementation and analysis of the results.
BibTeX for Cosmic topology IIIb
% Abbreviated author list; see the DOI or INSPIRE-HEP for the full record. @article{Samandar:2025iiib, author = {Samandar, Amirhossein and {Carr{\'o}n Duque}, Javier and Copi, Craig J. and {Martin Barandiaran}, Mikel and Mihaylov, Deyan P. and Starkman, Glenn D. and Akrami, Yashar and Anselmi, Stefano and Cornet-Gomez, Fernando and Eskilt, Johannes R. and others}, title = {Cosmic topology. {Part IIIb}. {Eigenmodes} and correlation matrices of spin-2 perturbations in orientable {Euclidean} manifolds}, journal = {Journal of Cosmology and Astroparticle Physics}, volume = {2025}, number = {08}, pages = {015}, year = {2025}, doi = {10.1088/1475-7516/2025/08/015}, eprint = {2503.08671}, archivePrefix = {arXiv}, primaryClass = {astro-ph.CO} } -
Co-author
Cosmic topology. Part Ic. Limits on lens spaces from circle searches
JCAP 01 (2025) 004 · 2025
Uses searches for matched circles in the CMB to set observational limits on lens-space topologies.
My contribution: Contributed to the manuscript text and preparation, and provided editorial suggestions.
BibTeX for Cosmic topology Ic
% Abbreviated author list; see the DOI or INSPIRE-HEP for the full record. @article{Saha:2025ic, author = {Saha, Samanta and Copi, Craig J. and Starkman, Glenn D. and Anselmi, Stefano and {Carr{\'o}n Duque}, Javier and {Martin Barandiaran}, Mikel and Akrami, Yashar and Cornet-Gomez, Fernando and Jaffe, Andrew H. and Kosowsky, Arthur and others}, title = {Cosmic topology. {Part Ic}. {Limits} on lens spaces from circle searches}, journal = {Journal of Cosmology and Astroparticle Physics}, volume = {2025}, number = {01}, pages = {004}, year = {2025}, doi = {10.1088/1475-7516/2025/01/004}, eprint = {2409.02226}, archivePrefix = {arXiv}, primaryClass = {astro-ph.CO} } -
First author Lead author
Cosmic topology. Part IIIa. Microwave background parity violation without parity-violating microphysics
JCAP 11 (2024) 020 · 2024
Shows that the shape of space alone, with no new parity-violating physics, can produce parity-odd patterns in CMB polarization.
Non-trivial spatial topology can induce parity-odd TB and EB correlations in the microwave background, an observational channel for both topology and primordial tensor modes.
My contribution: Led the analytical derivations, numerical implementation, and analysis of the results on topology-induced CMB parity violation.
BibTeX for Cosmic topology IIIa
% Abbreviated author list; see the DOI or INSPIRE-HEP for the full record. @article{Samandar:2024iiia, author = {Samandar, Amirhossein and {Carr{\'o}n Duque}, Javier and Copi, Craig J. and {Martin Barandiaran}, Mikel and Mihaylov, Deyan P. and Pereira, Thiago S. and Starkman, Glenn D. and Akrami, Yashar and Anselmi, Stefano and Cornet-Gomez, Fernando and others}, title = {Cosmic topology. {Part IIIa}. {Microwave} background parity violation without parity-violating microphysics}, journal = {Journal of Cosmology and Astroparticle Physics}, volume = {2024}, number = {11}, pages = {020}, year = {2024}, doi = {10.1088/1475-7516/2024/11/020}, eprint = {2407.09400}, archivePrefix = {arXiv}, primaryClass = {astro-ph.CO} } -
Co-author
Cosmic topology. Part IVa. Classification of manifolds using machine learning: a case study with small toroidal universes
JCAP 09 (2024) 057 · 2024
Tests whether machine-learning classifiers can tell small toroidal universes apart from CMB maps, in a setting where the optimal answer can also be computed.
Classifiers trained on simulated CMB maps approach likelihood-level discrimination at a fraction of the cost; because the likelihood is computable here, the learned model's gap to the optimal decision rule can be measured.
My contribution: Contributed to benchmarking learned topology classification against a directly computable likelihood, and helped prepare and revise the manuscript.
BibTeX for Cosmic topology IVa
% Abbreviated author list; see the DOI or INSPIRE-HEP for the full record. @article{Tamosiunas:2024iva, author = {Tamosiunas, Andrius and Cornet-Gomez, Fernando and Akrami, Yashar and Anselmi, Stefano and {Carr{\'o}n Duque}, Javier and Copi, Craig J. and Eskilt, Johannes R. and G{\"u}ng{\"o}r, {\"O}zen{\c{c}} and Jaffe, Andrew H. and Kosowsky, Arthur and others}, title = {Cosmic topology. {Part IVa}. {Classification} of manifolds using machine learning: a case study with small toroidal universes}, journal = {Journal of Cosmology and Astroparticle Physics}, volume = {2024}, number = {09}, pages = {057}, year = {2024}, doi = {10.1088/1475-7516/2024/09/057}, eprint = {2404.01236}, archivePrefix = {arXiv}, primaryClass = {astro-ph.CO} }
Accepted
-
Sixth author Accepted
Test-Time Scaling in Reasoning LLMs: Inference Regimes, Evaluation, and Reproducibility
Accepted for publication in TMLR · 2026
Sets out a common framework and reproducibility standards for test-time scaling, that is, spending more computation at inference time to improve a reasoning model's answers.
Formalizes test-time scaling as budgeted inference over the implicit prefix tree of an autoregressive transformer (sequential, leaf-level and prefix-level scaling), distinguishes exact replay from distributional reproducibility, and releases over 2 billion reasoning traces.
My contribution: Contributed to the discussion of LLM evaluation, the methods for assessing uncertainty, and the development of the new version of Scorio.
BibTeX for Test-time scaling
% Abbreviated author list; see arXiv for the full record. @article{Hariri:2026tts, author = {Hariri, Mohsen and Chen, Weicong and Shahini, Nahal and Singh, Vikash and Ye, K. and Samandar, Amirhossein and Ganguly, D. and others}, title = {Test-Time Scaling in Reasoning {LLMs}: Inference Regimes, Evaluation, and Reproducibility}, journal = {Transactions on Machine Learning Research}, note = {Accepted}, year = {2026}, eprint = {2608.04001}, archivePrefix = {arXiv}, primaryClass = {cs.LG} }
Preprints
Posted on arXiv; journal status is shown where known.
-
First author Lead authorPreprint
Beyond the Kullback–Leibler divergence: the likelihood-ratio distribution as a detectability measure
arXiv preprint · Submitted to arXiv 12 Oct 2026
Shows that the KL divergence, the standard forecast of whether a model is detectable, does not by itself give the probability of detecting it.
For multivariate Gaussian data, derives the distribution of the log-likelihood ratio and its cumulants under both hypotheses, and examines discovery potential and parameter-estimation accuracy when the null is a point and the alternative a continuous family. Case studies: a toy Gaussian model, the cubic three-torus, the tensor-to-scalar ratio from B-mode polarization, and dark energy from supernova distances. Detectability also depends on the size of the parameter space, the likelihood-ratio distributions, and the distance from parameter-space boundaries (the look-elsewhere effect), so the KL–detectability relation has to be calibrated for each model family and test statistic.
My contribution: Conceived the project and carried out the full analysis, including the analytical derivations, numerical implementation, and interpretation of the results.
BibTeX for Beyond the KL divergence
@article{Samandar:2026beyondKL, author = {Samandar, Amirhossein and {Martin Barandiaran}, Mikel and Lu, Linn Htat and Jaffe, Andrew H. and Mihaylov, Deyan P. and Starkman, Glenn D. and Trendafilova, C. and Akrami, Yashar and Anselmi, Stefano and {Carr{\'o}n Duque}, Javier and Copi, Craig J. and Cornet-Gomez, Fernando and Kosowsky, Arthur and Negro, Anna and Rodrigues, R. G. and Tamosiunas, Andrius}, title = {Beyond the {Kullback--Leibler} divergence: the likelihood-ratio distribution as a detectability measure}, year = {2026}, eprint = {2610.17186}, archivePrefix = {arXiv}, primaryClass = {astro-ph.CO} } -
Co-author Preprint
Cosmic topology. Part IId. Eigenmodes and correlation matrices of lens spaces
arXiv preprint · 2026
Computes the eigenmodes and correlation matrices needed to predict the CMB in lens-space universes.
My contribution: Contributed to the manuscript text and preparation, and provided editorial suggestions.
BibTeX for Cosmic topology IId
% Abbreviated author list; see arXiv for the full record. @article{MartinBarandiaran:2026iid, author = {{Martin Barandiaran}, Mikel and Petretti, Catherine and Anselmi, Stefano and Mihaylov, Deyan P. and Negro, Anna and others}, title = {Cosmic topology. {Part IId}. {Eigenmodes} and correlation matrices of lens spaces}, year = {2026}, eprint = {2610.12418}, archivePrefix = {arXiv} } -
Co-author Preprint
Cosmic topology. Part Va. Information content of the observable Universe
arXiv preprint · 2026
Asks how much information about the shape of space the observable Universe contains, comparing the CMB with the three-dimensional matter distribution.
Computes covariances of perturbation modes in the observed volume for the cubic three-torus (E1) and the three-torus with a quarter turn (E3).
My contribution: Contributed to the manuscript text and preparation, and provided editorial suggestions.
BibTeX for Cosmic topology Va
% Abbreviated author list; see arXiv for the full record. @article{Mihaylov:2026va, author = {Mihaylov, Deyan P. and {Martin Barandiaran}, Mikel and Anselmi, Stefano and {Carr{\'o}n Duque}, Javier and Starkman, Glenn D. and others}, title = {Cosmic topology. {Part Va}. {Information} content of the observable {Universe}}, year = {2026}, eprint = {2608.25033}, archivePrefix = {arXiv}, primaryClass = {astro-ph.CO} } -
Co-author Preprint
Quantum Signatures of Cosmic Topology: How Casimir Backreaction Transmits Isotropy Violation
arXiv preprint · 2026
Shows how quantum effects in a universe with non-trivial topology can make its expansion direction-dependent.
Quantum fields in topologically non-trivial universes carry a finite, scheme-independent Casimir stress-energy; its backreaction during de Sitter expansion drives anisotropic expansion from a locally isotropic start.
My contribution: Contributed to the manuscript text and preparation, and provided editorial suggestions.
BibTeX for Quantum signatures of cosmic topology
% Abbreviated author list; see arXiv for the full record. @article{Negro:2026casimir, author = {Negro, Anna and Hinterbichler, Kurt and Starkman, Glenn D. and Akrami, Yashar and Anselmi, Stefano and others}, title = {Quantum Signatures of Cosmic Topology: How {Casimir} Backreaction Transmits Isotropy Violation}, year = {2026}, eprint = {2603.12319}, archivePrefix = {arXiv}, primaryClass = {hep-th} } -
Co-author Under review at JCAP
Cosmic topology. Part IIc. Detectability with non-standard primordial power spectrum
arXiv preprint · 2026
Re-examines how detectable cosmic topology is when the primordial power spectrum departs from its standard form.
My contribution: Contributed to the manuscript text and preparation, and provided editorial suggestions.
BibTeX for Cosmic topology IIc
% Abbreviated author list; see arXiv for the full record. @article{Noltmann:2026iic, author = {Noltmann, Joline and Tamosiunas, Andrius and Mihaylov, Deyan P. and Akrami, Yashar and {Carr{\'o}n Duque}, Javier and others}, title = {Cosmic topology. {Part IIc}. {Detectability} with non-standard primordial power spectrum}, year = {2026}, eprint = {2602.15527}, archivePrefix = {arXiv}, primaryClass = {astro-ph.CO} }
In preparation
Not counted in the totals above.
-
First author Lead authorIn preparation
Cosmic topology. Part IIIc. Eigenmodes and correlation matrices of spin-2 perturbations in non-orientable Euclidean manifolds
In preparation
Extends the tensor-mode (spin-2) eigenmode and covariance calculation of Part IIIb to the non-orientable Euclidean manifolds.
My contribution: First author.
Software
-
CMBtopology
v0.1.0 (2025) · Python · CUDA · HPC
GPU-parallelized Python/CUDA/HPC package that computes CMB covariance matrices for compact Euclidean topologies; adopted across the COMPACT Collaboration for tensor-mode likelihood analysis.
My contribution: Developed the GPU-parallelized Python/HPC implementation.
-
Scorio
v0.2.2 · Python · Julia
Open-source Bayesian evaluation toolkit for LLMs and the official implementation of Bayes@N: Dirichlet-posterior estimates and credible intervals in place of Pass@k.
My contribution: Contributed to the development of the new version.