Search

Manual code reviews are an essential but time-consuming part of software development, often leading reviewers to prioritize technical issues while skipping valuable assessments. This paper presents an algorithmic model that automates aspects of code review typically avoided due to their complexity or subjectivity, such as assessing coding time, implementation time, and code complexity. Instead of replacing manual reviews, our model adds insights that help reviewers focus on more impactful tasks. Calibrated using expert evaluations, the model predicts key metrics from code commits with strong correlations to human judgments (r = 0.82 for coding time, r = 0.86 for implementation time). By automating these assessments, we reduce the burden on human reviewers and ensure consistent analysis of time-consuming areas, offering a scalable solution alongside manual reviews. This research shows how automated tools can enhance code reviews by addressing overlooked tasks, supporting data-driven decisions and improving the review process.

The light rays trajectories in Kerr metric, resulting from Fermat's principle, are considered from the point of view of integrable systems. It is shown how the counterpart of Carter constant emerges as a result of coupling-constant metamorphosis. The latter provides a convenient method of describing the null geodesics in Kerr metric., Comment: 11 pages, no figures Sereval misprints corrected; the number of arbitrary parameters in the Hamiltonian reduced from 3 to 1

Can large multimodal models have a human-like ability for emotional and social reasoning, and if so, how does it work? Recent research has discovered emergent theory-of-mind (ToM) reasoning capabilities in large language models (LLMs). LLMs can reason about people's mental states by solving various text-based ToM tasks that ask questions about the actors' ToM (e.g., human belief, desire, intention). However, human reasoning in the wild is often grounded in dynamic scenes across time. Thus, we consider videos a new medium for examining spatio-temporal ToM reasoning ability. Specifically, we ask explicit probing questions about videos with abundant social and emotional reasoning content. We develop a pipeline for multimodal LLM for ToM reasoning using video and text. We also enable explicit ToM reasoning by retrieving key frames for answering a ToM question, which reveals how multimodal LLMs reason about ToM.

The Fermat principle is advocated to be a convenient tool to analyze the light propagation in a curved space time. It is shown that in the weak deflection regime the light ray trajectories can be systematically described by applying the Lindstedt--Poicare method of solving perturbatively the nonlinear oscillation equations. The expansion in terms of inverse invariant impact parameter for Schwarzschild, Reissner--Nordstr\"om and Kerr (equatorial motion) metrics is described. The corresponding deflection angles are computed to the third order. Only algebraic operations are involved in the derivation; no integrations or Fourier expansion of elliptic functions are necessary., Comment: 27 pages, 13 figures, new version considerably extended

We introduce an approach to evaluate language model (LM) agency using negotiation games. This approach better reflects real-world use cases and addresses some of the shortcomings of alternative LM benchmarks. Negotiation games enable us to study multi-turn, and cross-model interactions, modulate complexity, and side-step accidental evaluation data leakage. We use our approach to test six widely used and publicly accessible LMs, evaluating performance and alignment in both self-play and cross-play settings. Noteworthy findings include: (i) only closed-source models tested here were able to complete these tasks; (ii) cooperative bargaining games proved to be most challenging to the models; and (iii) even the most powerful models sometimes "lose" to weaker opponents, Comment: Accepted to ICLR 2024, code and link to project data are made available at https://github.com/epfl-dlab/LAMEN

New form of Fermat's principle for light propagation in arbitrary (i.e. in general neither static nor stationary) gravitational field is proposed. It is based on Herglotz extension of canonical formalism and simple relation between the dynamics described by the Lagrangians homogeneous in velocities and the reduced dynamics on lower-dimensional configuration manifold. This approach is more flexible as it allows to extend immediately the Fermat principle to the case of massive particles and to eliminate any space-time coordinate, not only $x^0$., Comment: 27 pages, 1 figure, the title changed, considerably extended version

Consider a holomorphic map $F: D \to G$ between two domains in ${\mathbb C}^N$. Let $\mathcal F$ denote a family of geodesics for the Kobayashi distance, such that $F$ acts as an isometry on each element of $\mathcal F$. This paper is dedicated to characterizing the scenarios in which the aforementioned condition implies that $F$ is a biholomorphism. Specifically, we establish this when $D$ is a complete hyperbolic domain, and $\mathcal F$ comprises all geodesic segments originating from a specific point. Another case is when $D$ and $G$ are $C^{2+\alpha}$-smooth bounded pseudoconvex domains, and $\mathcal F$ consists of all geodesic rays converging at a designated boundary point of $D$. Furthermore, we provide examples to demonstrate that these assumptions are essentially optimal.

Studying the behavior of real and complex geodesics we provide sharp estimates for the Kobayashi distance, the Lempert function, and the Carath\'eodory distance on $\mathcal{C}^{2,\alpha}$-smooth strongly pseudoconvex domains. Similar estimates are also provided for the Bergman distance on strongly pseudoconvex domains with $\mathcal{C}^{3,1}$-boundary.

For vanishing fidelity between initial and final states two important quantum speed limits, the Mandelstam-Tamm limit (involving energy dispersion) and Margolus-Levitin one (involving excitation energy expectation value) have been derived. While the generalization of the former limit to the case of arbitrary fidelity is straightforward, the relevant generalization of the latter, given in the seminal paper by Giovanetti et al (Phys. Rev. A67 (2003), 052109) was based on the conjectured equality of lower and upper bounds on the right hand side of generalized Margolus-Levitin inequality, verified numerically up to seven digits. Only recently there appear two proofs of the conjecture. We provide below a very elementary new proof, based on the simplest tools from differential calculus. Thus the generalized Margolus-Levitin speed limit can be derived much in the spirit of the original one valid for vanishing fidelity., Comment: 14 pages, 8 figures, substantially revised according to the suggestions of the referees, published version

We report on the first production of an antihydrogen beam by charge exchange of 6.1 keV antiprotons with a cloud of positronium in the GBAR experiment at CERN. The antiproton beam was delivered by the AD/ELENA facility. The positronium target was produced from a positron beam itself obtained from an electron linear accelerator. We observe an excess over background indicating antihydrogen production with a significance of 3-4 standard deviations.

Carefully standardized facial images of 591 participants were taken in the laboratory, while controlling for self-presentation, facial expression, head orientation, and image properties. They were presented to human raters and a facial recognition algorithm: both humans (r=.21) and the algorithm (r=.22) could predict participants' scores on a political orientation scale (Cronbach's alpha=.94) decorrelated with age, gender, and ethnicity. These effects are on par with how well job interviews predict job success, or alcohol drives aggressiveness. Algorithm's predictive accuracy was even higher (r=.31) when it leveraged information on participants' age, gender, and ethnicity. Moreover, the associations between facial appearance and political orientation seem to generalize beyond our sample: The predictive model derived from standardized images (while controlling for age, gender, and ethnicity) could predict political orientation (r=.13) from naturalistic images of 3,401 politicians from the U.S., UK, and Canada. The analysis of facial features associated with political orientation revealed that conservatives tended to have larger lower faces. The predictability of political orientation from standardized images has critical implications for privacy, the regulation of facial recognition technology, and understanding the origins and consequences of political orientation., Comment: Now published at American Psychologist (2024)

We describe a general procedure which allows to construct, starting from a given Hamiltonian, the whole family of new ones sharing the same set of unparameterized trajectories in phase space. The symmetry structure of this family can be completely characterized provided the symmetries of initial Hamiltonian are known. Our approach covers numerous models considered in literature as well as it allows to construct novel ones.. It provides a far reaching generalization of Hietarinta et al. coupling-constant metamorphosis method and another proof of Darboux theorem., Comment: 25 pages; no figures; considerably extended new version

We prove that in a strongly pseudoconvex domain with smooth boundary, then the length of a geodesic for the Kobayashi-Royden infinitesimal metric between two points is bounded by a constant multiple of the Euclidean distance between the points.

Eleven Large Language Models (LLMs) were assessed using a custom-made battery of false-belief tasks, considered a gold standard in testing Theory of Mind (ToM) in humans. The battery included 640 prompts spread across 40 diverse tasks, each one including a false-belief scenario, three closely matched true-belief control scenarios, and the reversed versions of all four. To solve a single task, a model needed to correctly answer 16 prompts across all eight scenarios. Smaller and older models solved no tasks; GPT-3-davinci-003 (from November 2022) and ChatGPT-3.5-turbo (from March 2023) solved 20% of the tasks; ChatGPT-4 (from June 2023) solved 75% of the tasks, matching the performance of six-year-old children observed in past studies. We explore the potential interpretation of these findings, including the intriguing possibility that ToM, previously considered exclusive to humans, may have spontaneously emerged as a byproduct of LLMs' improving language skills., Comment: TRY RUNNING ToM EXPERIMENTS ON YOUR OWN: The code and tasks used in this study are available at Colab (https://colab.research.google.com/drive/1ZRtmw87CdA4xp24DNS_Ik_uA2ypaRnoU). Don't worry if you are not an expert coder, you should be able to run this code with no-to-minimum Python skills. Or copy-paste the tasks to ChatGPT's web interface

We characterize polynomials that are cyclic in Dirichlet-type spaces in the unit ball in $\mathbb C^2$

Large language models (LLMs) are currently at the forefront of intertwining AI systems with human communication and everyday life. Therefore, it is of great importance to evaluate their emerging abilities. In this study, we show that LLMs like GPT-3 exhibit behavior that strikingly resembles human-like intuition - and the cognitive errors that come with it. However, LLMs with higher cognitive capabilities, in particular ChatGPT and GPT-4, learned to avoid succumbing to these errors and perform in a hyperrational manner. For our experiments, we probe LLMs with the Cognitive Reflection Test (CRT) as well as semantic illusions that were originally designed to investigate intuitive decision-making in humans. Our study demonstrates that investigating LLMs with methods from psychology has the potential to reveal otherwise unknown emergent traits.

It is proved for a strongly pseudoconvex domain $D$ in $\Bbb C^d$ with $\mathcal C^{2,\alpha}$-smooth boundary that any complex geodesic through every two close points of $D$ sufficiently close to $\partial D$ and whose difference is non-tangential to $\partial D$ intersect a compact subset of $D$ that depends only on the rate of non-tangentiality. As an application, a lower bound for the Kobayashi distance is obtained., Comment: v2: to appear in Math. Ann

It is shown that by an appropriate canonical transformation Kepler dynamics can be put in the form which allows to exhibit the structure of the symmetry transformations related to the superintegrability. They appear to fit nicely into general scheme of nonlinear realizations. In new coordiantes the Kepler dynamics results from dimensional reduction of that describing low energy mesons with spontaneously broken chiral symmetry., Comment: 16 pages, no figures; revised version to appear in Physical Review E

We generalize the results obtained recently (Nonlinearity \underline{36} (2023), 1143) by providing a very simple proof of the superintegrability of the Hamiltonian $H=\vec{p}\,^{2}+F(\vec{q}\cdot\vec{p})$, $\vec{q}, \vec{p}\in\mathbb{R}^{2}$, for any analytic function $F$. The additional integral of motion is constructed explicitly and shown to reduce to a polynomial in canonical variables for polynomial $F$. The generalization to the case $\vec{q}, \vec{p}\in \mathbb{R}^{n}$ is sketched., Comment: 9 pages, no figures; pasections reorganized, references updated, some misprints corrected

The nongeneric six- and eightdimensional orbits of SO(4,2) are described in explicitly covariant way. The relevant Hamiltonian dynamical systems are constructed and canonically quantized. It is shown that the resulting unitary representations of conformal group fit into the classification described by Mack (G. Mack, Comm. Math. Phys. 55 (1977),1)., Comment: 20 pages, no figures

We study the boundedness of composition operators on the weighted Bergman spaces and the Hardy space over the polydisc. For arbitrary polydisc we prove the rank sufficiency theorem which, in particular, provides us with a simple criterion describing boundedness of composition operators on the spaces over the bidisc. Such a consistent characterization is obtained for the classical Bergman space over the tridisc.

We present a description of the GBAR positron (e+) trapping apparatus, which consists of a three stage Buffer Gas Trap (BGT) followed by a High Field Penning Trap (HFT), and discuss its performance. The overall goal of the GBAR experiment is to measure the acceleration of the neutral antihydrogen (H) atom in the terrestrial gravitational field by neutralising a positive antihydrogen ion (H+), which has been cooled to a low temperature, and observing the subsequent H annihilation following free fall. To produce one H+ ion, about 10^10 positrons, efficiently converted into positronium (Ps), together with about 10^7 antiprotons (p), are required. The positrons, produced from an electron linac-based system, are accumulated first in the BGT whereafter they are stacked in the ultra-high vacuum HFT, where we have been able to trap 1.4(2) x 10^9 positrons in 1100 seconds.

We show that for every connected analytic subvariety $V$ there is a pseudoconvex set $\Omega$ such that every bounded matrix-valued holomorphic function on $V$ extends isometrically to $\Omega$. We prove that if $V$ is two analytic disks intersecting at one point, if every bounded scalar valued holomorphic function extends isometrically to $\Omega$, then so does every matrix-valued function. In the special case that $\Omega$ is the symmetrized bidisk, we show that this cannot be done by finding a linear isometric extension from the functions that vanish at one point.

If $V$ is an analytic set in a pseudoconvex domain $\Omega$, we show there is always a pseudoconvex domain $G \subseteq \Omega$ that contains $V$ and has the property that every bounded holomorphic function on $V$ extends to a bounded holomorphic function on $G$ with the same norm. We find such a $G$ for some particular analytic sets. When $\Omega$ is an operhedron we show there is a norm on holomorphic functions on $V$ that can always be preserved by extensions to $\Omega$.

The stability of the spin polaron quasiparticle, well established in studies of a single hole in the 2D antiferromagnets, is investigated in the 1D antiferromagnets using a t-J model. We perform an exact slave fermion transformation to the holon-magnon basis, and diagonalize numerically the resulting model in the presence of a single hole. We demonstrate that the spin polaron collapses - and the spin-charge separation takes over - due to the specific role played by the magnon-magnon interactions and the magnon hard-core constraint in the 1D t-J model. Moreover, we prove that the spin polaron is stable for any strength of the magnon-magnon interaction other than the unique value found in a 1D antiferromagnet with the continuous symmetry of the spin interactions. Fine-tuning to this unique value is extremely unlikely to occur in quasi-1D antiferromagnets, therefore the spin polaron is the stable quasiparticle of realistic 1D materials. Our results lead to a new interpretation of the ARPES spectra of quasi-1D antiferromagnets in the spin polaron language., Comment: 26 pages, 13 figures; revised version resubmitted to SciPost in June 2024 (Sec. 5 and App. C added)