arXiv subjects

Aadarsh Singh

Publications and source records attributed to Aadarsh Singh.

11 recordsLinked to original sources

arXiv2026

HEPToolBench 1.2: Testing How Reliably Language Models Can Drive Particle Physics Software

Scientists increasingly want to drive research software by natural-language request, but fluent output helps only if it becomes a correct machine-readable artifact. We introduce HEPToolBench, a benchmark of 28 collider-simulation tasks scored by deterministic, task-specific scorers, plus a three-task structured-debugging extension. We evaluate 42 deployments, from small locally served open-weight models to hosted frontier systems. The central experiment compares direct generation of native HEP-tool syntax with a schema-mediated interface where the model returns a typed representation that deterministic software serializes. Across five matched requests, the mean score rises from 0.418 to 0.902 and task passes from 21/210 to 159/210, with 41 of 42 deployments improving; eleven deployments, seven locally served open-weight models, go from 0/5 passes under native syntax to 5/5 under the structured interface. Values use the corrected v1.2.1 contract, in which the two native scorers that had enforced operational conventions absent from their own prompts are rescored prompt-faithfully cohort-wide. One asymmetry remains: native scorers score the recorded response verbatim and reject Markdown-fenced replies, whereas structured scorers recover JSON from such wrappers; no full-cohort rescoring under a single symmetric extraction rule has been done. An audit of archived responses from 17 locally served deployments shows this does not explain their gains: passes remain six to ten times more frequent under the structured interface when both sides use the same extraction rule. A task pass alone does not guarantee runtime or scientific viability. Within this scope, moving syntax generation into deterministic software can substantially improve reliability for both small local and frontier models. Prompts, scorers, responses, and regeneration scripts are released for independent evaluation and extension.

hep-ph↗

arXiv2026

HEPLocalAgent 1.0: Running Collider Simulations from Plain-Language Requests on Your Own Computer

We present HEPLocalAgent, an open source local interface that builds a bounded class of collider simulation workflows from natural language requests. A locally served language model proposes a typed workflow representation, and deterministic software then restores recognized user stated quantities, builds the HEP tool inputs, validates the supported workflow, and presents the artifacts for approval before execution. In a same response comparison on 47 evaluable model request cases, the first structured proposal gave 7 unmodified artifacts satisfying the external benchmark scorer, against 19 after the full deterministic pipeline. Under the fixed representation normalization defined by the benchmark, the counts were 11 and 43. The direction of improvement is unchanged. The gap between the two views arises because the released builder and the benchmark scorers disagree on three bookkeeping conventions, namely launch form, two fixed control lines, and the output directory name, not on physics content. Four of seven approved workflows ran to completion on the managed local software stack, with cross sections consistent between repeats. In a separate challenge set, 57 of 96 problematic requests still reached the approval stage after part of the request was dropped, defaulted, or reinterpreted. No tested unsafe payload was retained in an executable artifact before the approval gate, but this does not establish operating system level containment. The deterministic backend supports MadGraph, Pythia8, Delphes, and a restricted MadAnalysis 5 plan. Reliable natural language routing to the MadAnalysis stage was not demonstrated in the tested examples. Version 1 should therefore be seen as an inspectable, validation gated workflow constructor requiring expert approval rather than an autonomous or scientifically self validating agent.

hep-ph↗

arXiv2026

Fourier-Preconditioned Path Deformations for Multi-Field Vacuum Tunnelling

We present an endpoint-safe Fourier method for multi-field vacuum tunnelling. The field-space tunnelling path is written as a straight-line interpolation between the false and true vacua, plus sine-mode deformations that vanish at the endpoints. This gives a finite-dimensional path optimisation problem, which we implement using automatic differentiation in the JAX numerical framework. The method is studied both as a standalone variational ansatz for curved tunnelling paths and as a preconditioner for existing bounce solvers. On the OptiBounce benchmark potential for $N_\phi=3,\ldots,20$ and on a nested random-coefficient potential family up to $N_\phi=50$, the Fourier result agrees with FindBounce, OptiBounce, and CosmoTransitions at the sub-percent level in the regular benchmark cases, while requiring only a modest number of modes. We also compare several endpoint-safe basis families and find that Fourier sine modes provide a robust default for smooth tunnelling paths. When used as an initialiser, the Fourier path supplies useful geometric information to existing solvers before the final bounce calculation is carried out. In the CosmoTransitions tests, this reduces the number of steps in subsequent path deformation, while in the FindBounce point-injection tests, it gives large runtime improvements in the high-dimensional cases up to 90 $\%$. These results suggest that endpoint-safe Fourier paths provide a useful bridge between simple analytic path ans\"atze and fully numerical multi-field bounce algorithms.

hep-ph↗

arXiv2026

Theoretical and Experimental Constraints in the $\mu$--$\tau$ Four-Lepton Sector of the SMEFT: implications to neutrino self interactions

We study the dimension-six SMEFT four-lepton operators in the $\mu$--$\tau$ sector. These operators control both charged-lepton scattering and neutrino self-interactions, the latter being weakly constrained by direct laboratory probes despite their importance for cosmological tensions. We compare three classes of constraints on the Warsaw-basis coefficients $[C_{\ell\ell}]_{2222}$, $[C_{\ell\ell}]_{2233}$, and $[C_{\ell\ell}]_{2332}$. We use perturbative unitarity from $2\!\to\!2$ partial-wave analysis, spin-summing positivity sum rules, and the experimental bounds from NA64$\mu$ and the global fit~\cite{Falkowski:2017pss}. We find that the global fit dominates for $[C_{\ell\ell}]_{2222}$ and $[C_{\ell\ell}]_{2332}$, while NA64$\mu$ provides the leading bound on $[C_{\ell\ell}]_{2233}$, with the unitarity line for this direction entering the range of collider energies near $200~\mathrm{GeV}$. Renormalization-group running between $1~\mathrm{GeV}$ and $30~\mathrm{TeV}$ modifies these coefficients by up to $10\%$. Translating these bounds onto the effective four-neutrino coupling $G_\mathrm{eff}$, we find values many orders of magnitude smaller than the strongly interacting regime motivated by the Hubble tension; this excludes heavy-mediator UV completions of strong $\nu_{\mu}$--$\nu_{\mu}$ and $\nu_\mu$--$\nu_\tau$ self-interactions within the validity of the dimension-six SMEFT and in the absence of tuned cancellations between operators, while leaving the cosmologically motivated light-mediator scenarios unconstrained by this analysis. Finally, we comment on the bounds these coefficients place on a leptophilic $L_\mu - L_\tau$ $Z'$ UV completion. Our SMEFT-based current and projected NA64$\mu$ bounds reproduce the dedicated $Z'$ analyses already available in the literature.

hep-ph↗

arXiv2025

Can Randomness lead to non-anarchical mixing angles ?

We revisit the proposal of Craig and Sutherland that Anderson localization in a disordered fermion theory space can generate small neutrino masses from TeV scale physics \citecraig2018exponential}. Building on this idea, we ask a broader question: can randomness in fermion mass parameters also give rise to nonanarchical neutrino mixing angles, and how does the answer depend on the geometry of the mass graph? To explore this, we analyse three representative geometries a nearest neighbour chain, a fully connected non local model, and the Petersen graph in both Dirac and Majorana neutrino realisations. In the regime of strong diagonal disorder, all geometries display robust localization and naturally generate the observed neutrino mass scale, with the corresponding flavour mixing angles reflecting the random localization centres and thus taking an anarchical form. In the regime of weak disorder, where localization is milder, and eigenmodes can exhibit quasidegeneracies, light neutrino masses can emerge through GIM-mechanismlike cancellations among the heavy states. The weak disorder with geometry dependent weak localization constitutes a distinct pathway to structured mixings within disordered theory spaces. Overall, our results delineate the regimes in which disorder driven mechanisms produce hierarchical masses and identify the conditions under which structured flavour mixing can arise.

hep-ph↗

arXiv2025

Flavour from Fractal Mass Chains

We explore the possibility that the underlying flavour structure of the Standard Model could be determined by mass chains on a fractal geometry. We consider, as an example, the theory space on a Sierpinski-like geometry. The fermion mass chains on a Sierpinski-like geometry with three decorations (iterations) lead to three zero modes, which can be identified with the three generations of the Standard Model. This framework also reproduces the measured charged and neutral lepton masses and mixing angles with very few parameters. We also briefly discuss the possible extension to the quark sector.

hep-ph↗

arXiv2024

Independent Chiral Control in Theory-Space Models:A Rank-Preserving Framework and Its Application to Neutrino Mass Generation

We develop a general framework of rank-preserving, element-wise matrix transformations for engineering fermion mass hierarchies in theory-space constructions. We prove that preservation of massless modes requires the transformation function to be separable, $g_f(i,j)=g^{(L)}_f(i)g^{(R)}_f(j)$, which in turn enables independent control of left- and right-chiral zero-mode profiles directly at the level of the theory-space mass matrix. This formalism unifies and extends the clockwork mechanism, permits controlled deformation of Kaluza--Klein spectra, and enhances hierarchy generation in GIM-like fine-cancellation scenarios. As a concrete application, we show that in theory-space models for neutrino masses, suitable transformations allow sub-eV light neutrinos to arise from TeV-scale new physics with only $\mathcal{O}(40)$ additional fermionic sites, while remaining consistent with charged-lepton flavor-violation bounds. In contrast, the corresponding untransformed models asymptote at the MeV scale and cannot access the phenomenologically required regime without extreme field multiplicities or hierarchical parameters.

math.RA↗

arXiv2024

Revisiting Neutrino Masses In Clockwork Models

In this paper, we have studied a mechanism that naturally produces hierarchical masses using fine cancellation rather than widely used suppression mechanisms as in seesaw, clockwork or randomness-assisted localization models. We have also looked at various variants of the clockwork model and provided analytical expressions for their 0 modes. Few generalizations of clockwork models such as generalized CW and next to nearest neighbour interaction CW have already been explored by a few authors. All the variants of CW models have the same underlying principle i.e., suppression. In this study, it was found that non-local CW models relaxes the $\left| q \right| > 1$ constraint of ordinary CW models to produce localization. We also made a comparison among them and have shown that in some scenarios variants of CW are more efficient than ordinary CW. Additionally, the fine-cancellation (precision-prune) mechanism is depicted within the framework of the extra dimension picture. Finally, some phenomenological signatures of all these models are also discussed along with their benchmark points.

hep-ph↗

arXiv2021

SciDr at SDU-2020: IDEAS -- Identifying and Disambiguating Everyday Acronyms for Scientific Domain

We present our systems submitted for the shared tasks of Acronym Identification (AI) and Acronym Disambiguation (AD) held under Workshop on SDU. We mainly experiment with BERT and SciBERT. In addition, we assess the effectiveness of "BIOless" tagging and blending along with the prowess of ensembling in AI. For AD, we formulate the problem as a span prediction task, experiment with different training techniques and also leverage the use of external data. Our systems rank 11th and 3rd in AI and AD tasks respectively.

cs.CL↗

arXiv2020

NutCracker at WNUT-2020 Task 2: Robustly Identifying Informative COVID-19 Tweets using Ensembling and Adversarial Training

We experiment with COVID-Twitter-BERT and RoBERTa models to identify informative COVID-19 tweets. We further experiment with adversarial training to make our models robust. The ensemble of COVID-Twitter-BERT and RoBERTa obtains a F1-score of 0.9096 (on the positive class) on the test data of WNUT-2020 Task 2 and ranks 1st on the leaderboard. The ensemble of the models trained using adversarial training also produces similar result.

cs.CL↗

arXiv2020

DSC IIT-ISM at SemEval-2020 Task 6: Boosting BERT with Dependencies for Definition Extraction

We explore the performance of Bidirectional Encoder Representations from Transformers (BERT) at definition extraction. We further propose a joint model of BERT and Text Level Graph Convolutional Network so as to incorporate dependencies into the model. Our proposed model produces better results than BERT and achieves comparable results to BERT with fine tuned language model in DeftEval (Task 6 of SemEval 2020), a shared task of classifying whether a sentence contains a definition or not (Subtask 1).

cs.CL↗