arXiv subjects

João Rodrigues

Publications and source records attributed to João Rodrigues.

At least 19 recordsLinked to original sources

arXiv2026

Non-perturbative data for Weil-Petersson volumes and intersection numbers using ordinary differential equations

Recently, a new method was introduced for computing $V_{g,1}(b)$, the Weil-Petersson volumes of the moduli space of Riemann surfaces of genus $g$ with one geodesic boundary of length $b$, various supersymmetric generalizations of them, as well as analogous quantities in intersection theory. The physical setting is the computation of a certain one-point function in a variety of models of 2D gravity for which there is a double-scaled random matrix model (RMM) description. The method combines perturbative solutions of two ordinary differential equations (ODEs), the Gel'fand-Dikii resolvent equation, and the RMM's string equation. In this paper, we extend the method to extract non-perturbative information about the $V_{g,1}(b)$ (and their analogues) that is naturally contained in the full ODEs, providing an efficient prescription for computing the transseries coefficients of the one-point correlation function, fully incorporating ZZ-brane and FZZT-brane effects, and for the first time, mixed ZZ-FZZT-effects. We use as a case study the (2,3) minimal string, computing perturbative and non-perturbative quantities, comparing them to perturbative results from topological recursion, and to results from the recent non-perturbative topological recursion framework. As a particularly powerful further application we provide general predictions for the large order in $g$ growth of $V_{g,1}(b)$, and apply them to JT gravity, finding agreement with known results, and for analogous quantities in ${N} {=} 1$ JT supergravity, proving a conjecture of Stanford and Witten. Our predictions yield new growth formulae for the cases of ${N} {=} 2$ and ${N}{=}4$ JT supergravity.

hep-th↗

arXiv2026

Asymptotic bootstrap for unitary matrix integrals at complex coupling

We apply an asymptotic bootstrap estimate method to the non-perturbative study of unitary matrix integrals. The method combines exact recursion relations with asymptotic control of large modes to achieve very high numerical precision without relying on positivity or semidefinite programming. We demonstrate its effectiveness in large-$N$ unitary matrix models by computing Wilson loop expectation values with sensitivity to exponentially small instanton effects and validating them against analytical instanton calculations. We further use the method to explore phase diagrams of unitary matrix models in complex 't Hooft coupling space, where positivity is absent, and observe that Stokes lines provide a useful proxy for additional phase boundaries. Our results show that asymptotic bootstrap estimates offer a practical and precise tool for probing the non-perturbative structure of unitary matrix integrals.

hep-th↗

arXiv2026

Exact Solutions to Matrix Models and String Theories: The Local Construction

Exact nonperturbative solutions to hermitian one-matrix models, their topological string duals, as well as their double-scaling limits to multicritical and minimal string theories, may be obtained via the use of resurgent transseries. These solutions are generically resonant, entailing both eigenvalues and anti-eigenvalues, or, equivalently, both D-branes and negative-tension D-branes -- but are otherwise intricate to write down, having been previously addressed on a case-by-case approach. This work shows how there is a general and rather compact way to write down all these exact and fully nonperturbative transseries solutions in closed-form, immediately starting from the spectral geometry of the matrix model or string theory at hand, in the form of a discrete Fourier or Zak transform for their partition functions. This structure is inherently associated to the existence of anti-eigenvalues or negative-tension D-branes. The validity of these solutions is testable across all values of the parameters -- from weak to strong 't Hooft coupling, from small to large N; equivalently, from semi-classical to deeply quantum regimes -- and many such nonperturbative tests are performed against diverse examples ranging from matrix models to non-critical strings, fully validating our analytical and exact expressions. In particular, anti-eigenvalues or negative-tension D-branes are absolutely required to find sharp numerical matches. In order to study these solutions globally across their phase diagrams, however, complete non-linear Stokes data is still needed -- which will be addressed in a complementary follow-up paper.

hep-th↗

arXiv2025

Greybody factors of string-corrected d-dimensional black holes

We compute analytically greybody factors for asymptotically flat spherically symmetric black holes with stringy higher derivative corrections in d dimensions in the high frequency limit. Our calculations include both the eikonal limit - where the real part of the frequency of the scattered wave is much larger than the imaginary part - and the highly damped case - where the imaginary part of the frequency is much larger than the real part -, addressing the emission of gravitons and test scalar fields, and yielding full transmission and reflection scattering coefficients.

hep-th↗

arXiv2025

Hands-off Image Editing: Language-guided Editing without any Task-specific Labeling, Masking or even Training

Instruction-guided image editing consists in taking an image and an instruction and deliverring that image altered according to that instruction. State-of-the-art approaches to this task suffer from the typical scaling up and domain adaptation hindrances related to supervision as they eventually resort to some kind of task-specific labelling, masking or training. We propose a novel approach that does without any such task-specific supervision and offers thus a better potential for improvement. Its assessment demonstrates that it is highly effective, achieving very competitive performance.

cs.CL↗

arXiv2024

Open Sentence Embeddings for Portuguese with the Serafim PT* encoders family

Sentence encoder encode the semantics of their input, enabling key downstream applications such as classification, clustering, or retrieval. In this paper, we present Serafim PT*, a family of open-source sentence encoders for Portuguese with various sizes, suited to different hardware/compute budgets. Each model exhibits state-of-the-art performance and is made openly available under a permissive license, allowing its use for both commercial and research purposes. Besides the sentence encoders, this paper contributes a systematic study and lessons learned concerning the selection criteria of learning objectives and parameters that support top-performing encoders.

cs.CL↗

arXiv2024

Meta-prompting Optimized Retrieval-augmented Generation

Retrieval-augmented generation resorts to content retrieved from external sources in order to leverage the performance of large language models in downstream tasks. The excessive volume of retrieved content, the possible dispersion of its parts, or their out of focus range may happen nevertheless to eventually have a detrimental rather than an incremental effect. To mitigate this issue and improve retrieval-augmented generation, we propose a method to refine the retrieved content before it is included in the prompt by resorting to meta-prompting optimization. Put to empirical test with the demanding multi-hop question answering task from the StrategyQA dataset, the evaluation results indicate that this method outperforms a similar retrieval-augmented system but without this method by over 30%.

cs.CL↗

arXiv2024

PORTULAN ExtraGLUE Datasets and Models: Kick-starting a Benchmark for the Neural Processing of Portuguese

Leveraging research on the neural modelling of Portuguese, we contribute a collection of datasets for an array of language processing tasks and a corresponding collection of fine-tuned neural language models on these downstream tasks. To align with mainstream benchmarks in the literature, originally developed in English, and to kick start their Portuguese counterparts, the datasets were machine-translated from English with a state-of-the-art translation engine. The resulting PORTULAN ExtraGLUE benchmark is a basis for research on Portuguese whose improvement can be pursued in future work. Similarly, the respective fine-tuned neural language models, developed with a low-rank adaptation approach, are made available as baselines that can stimulate future work on the neural processing of Portuguese. All datasets and models have been developed and are made available for two variants of Portuguese: European and Brazilian.

cs.CL↗

arXiv2024

Advancing Generative AI for Portuguese with Open Decoder Gervásio PT*

To advance the neural decoding of Portuguese, in this paper we present a fully open Transformer-based, instruction-tuned decoder model that sets a new state of the art in this respect. To develop this decoder, which we named Gervásio PT*, a strong LLaMA~2 7B model was used as a starting point, and its further improvement through additional training was done over language resources that include new instruction data sets of Portuguese prepared for this purpose, which are also contributed in this paper. All versions of Gervásio are open source and distributed for free under an open license, including for either research or commercial usage, and can be run on consumer-grade hardware, thus seeking to contribute to the advancement of research and innovation in language technology for Portuguese.

cs.CL↗

arXiv2024

Fostering the Ecosystem of Open Neural Encoders for Portuguese with Albertina PT* Family

To foster the neural encoding of Portuguese, this paper contributes foundation encoder models that represent an expansion of the still very scarce ecosystem of large language models specifically developed for this language that are fully open, in the sense that they are open source and openly distributed for free under an open license for any purpose, thus including research and commercial usages. Like most languages other than English, Portuguese is low-resourced in terms of these foundational language resources, there being the inaugural 900 million parameter Albertina and 335 million Bertimbau. Taking this couple of models as an inaugural set, we present the extension of the ecosystem of state-of-the-art open encoders for Portuguese with a larger, top performance-driven model with 1.5 billion parameters, and a smaller, efficiency-driven model with 100 million parameters. While achieving this primary goal, further results that are relevant for this ecosystem were obtained as well, namely new datasets for Portuguese based on the SuperGLUE benchmark, which we also distribute openly.

cs.CL↗

arXiv2023

WKB method and quasinormal modes of string-theoretical d-dimensional black holes

After a brief introduction to quasinormal modes in dissipative systems, we review the WKB formalism in the context of the analytical calculation of quasinormal frequencies. We apply these results to the calculation of quasinormal frequencies associated with gravitational perturbations of d-dimensional spherically symmetric black holes with string corrections. We do this for two distinct limits: the eikonal limit and the asymptotic limit.

gr-qc↗

arXiv2023

The isospectrality of asymptotic quasinormal modes of large Gauss-Bonnet d-dimensional black holes

We compute the quasinormal frequencies of d-dimensional large spherically symmetric black holes with Gauss-Bonnet corrections in the highly damped regime. We solve perturbatively the master differential equation and we compute the monodromies of the master perturbation variable (analytically continued to the complex plane) in different contours, in order to obtain the quasinormal mode spectra. We consider tensorial, vectorial and scalar gravitational perturbations, obtaining the same frequencies for the three cases like in Einstein gravity. We also separately perform the same calculation for test scalar fields.

hep-th↗

arXiv2023

Advancing Neural Encoding of Portuguese with Transformer Albertina PT-*

To advance the neural encoding of Portuguese (PT), and a fortiori the technological preparation of this language for the digital age, we developed a Transformer-based foundation model that sets a new state of the art in this respect for two of its variants, namely European Portuguese from Portugal (PT-PT) and American Portuguese from Brazil (PT-BR). To develop this encoder, which we named Albertina PT-*, a strong model was used as a starting point, DeBERTa, and its pre-training was done over data sets of Portuguese, namely over data sets we gathered for PT-PT and PT-BR, and over the brWaC corpus for PT-BR. The performance of Albertina and competing models was assessed by evaluating them on prominent downstream language processing tasks adapted for Portuguese. Both Albertina PT-PT and PT-BR versions are distributed free of charge and under the most permissive license possible and can be run on consumer-grade hardware, thus seeking to contribute to the advancement of research and innovation in language technology for Portuguese.

cs.CL↗

arXiv2023

Character Simulation Using Imitation Learning With Game Engine Physics

Creating visual 3D sensing characters that interact with AI peers and virtual environments can be a difficult task for those with less experience in using learning algorithms or creating visual environments to execute an agent-based simulation. In this paper, the use of game engines as a tool to create and execute graphic simulations with 3D sensing characters is being explored with plugins such as ML-Agents for the Unity3D game engine. This allows the simulation of agents using off-the-shelf algorithms and using the game engine's motor for the visualizations of these agents. We explore the use of these tools to create visual bots for games, and teach them how to play the game until they reach a level where they can serve as adversaries for real-life players in interactive games.

cs.GR↗

arXiv2022

Transfer Learning of Lexical Semantic Families for Argumentative Discourse Units Identification

Argument mining tasks require an informed range of low to high complexity linguistic phenomena and commonsense knowledge. Previous work has shown that pre-trained language models are highly effective at encoding syntactic and semantic linguistic phenomena when applied with transfer learning techniques and built on different pre-training objectives. It remains an issue of how much the existing pre-trained language models encompass the complexity of argument mining tasks. We rely on experimentation to shed light on how language models obtained from different lexical semantic families leverage the performance of the identification of argumentative discourse units task. Experimental results show that transfer learning techniques are beneficial to the task and that current methods may be insufficient to leverage commonsense knowledge from different lexical semantic families.

cs.CL↗

arXiv2021

Asymptotic quasinormal modes of string-theoretical $d$-dimensional black holes

We compute the quasinormal frequencies of $d$-dimensional spherically symmetric black holes with leading string $α'$ corrections for tensorial gravitational perturbations in the highly damped regime. We solve perturbatively the master differential equation and we compute the monodromies of the master perturbation variable (analytically continued to the complex plane) in different contours, in order to obtain the quasinormal mode spectra. We proceed analogously for the quasinormal modes of test scalar fields. Differently than in Einstein gravity, we obtain distinct results for the two cases.

hep-th↗

arXiv2021

Eikonal quasinormal modes and shadow of string-corrected $d$-dimensional black holes

We compute the quasinormal frequencies of $d$-dimensional spherically symmetric black holes with leading string $α'$ corrections in the eikonal limit for tensorial gravitational perturbations and scalar test fields. We find that, differently than in Einstein gravity, the real parts of the frequency are no longer equal for these two cases. The corresponding imaginary parts remain equal to the principal Lyapunov exponent corresponding to circular null geodesics, to first order in $α'$. We also compute the radius of the shadow cast by these black holes.

hep-th↗

arXiv2020

Comparative Probing of Lexical Semantics Theories for Cognitive Plausibility and Technological Usefulness

Lexical semantics theories differ in advocating that the meaning of words is represented as an inference graph, a feature mapping or a vector space, thus raising the question: is it the case that one of these approaches is superior to the others in representing lexical semantics appropriately? Or in its non antagonistic counterpart: could there be a unified account of lexical semantics where these approaches seamlessly emerge as (partial) renderings of (different) aspects of a core semantic knowledge base? In this paper, we contribute to these research questions with a number of experiments that systematically probe different lexical semantics theories for their levels of cognitive plausibility and of technological usefulness. The empirical findings obtained from these experiments advance our insight on lexical semantics as the feature-based approach emerges as superior to the other ones, and arguably also move us closer to finding answers to the research questions above.

cs.CL↗