Despite impressive high-level video comprehension, multimodal language models struggle with spatial reasoning across time and space. While current spatial training approaches rely on real-world video data, obtaining diverse footage with precise spatial annotations remains a bottleneck. To alleviate this bottleneck, we present SIMS-V -- a systematic data-generation framework that leverages the privileged information of 3D simulators to create spatially-rich video training data for multimodal language models. Using this framework, we investigate which properties of simulated data drive effective real-world transfer through systematic ablations of question types, mixes, and scales. We identify a minimal set of three question categories (metric measurement, perspective-dependent reasoning, and temporal tracking) that prove most effective for developing transferable spatial intelligence, outperforming comprehensive coverage despite using fewer question types. These insights enable highly efficient training: our 7B-parameter video LLM fine-tuned on just 25K simulated examples outperforms the larger 72B baseline and achieves competitive performance with proprietary models on rigorous real
The deal takes the maker of titles including The Sims and EA FC into private ownership - and loads it with debt
We give inductive constructions of independent graphs that contain implied nonedges but do not contain any non-trivial rigid subgraphs, or \emph{nucleations}: some of the constructions and proofs apply to 3-dimensional abstract rigidity matroids with their respective definitions of nucleations and implied nonedges. The first motivation for the inductive constructions of this paper, which generate an especially intractable class of flexible circuits, is to illuminate further obstacles to settling Graver's maximality conjecture that the 3-dimensional generic rigidity matroid is isomorphic to Whiteley's cofactor matroid (the unique maximal matroid in which all graphs isomorphic to $K_5$ are circuits). While none of the explicit examples we provide refutes the maximality conjecture (since their properties hold in both matroids) the construction schemes are useful regardless whether the conjecture is true or false, e.g. for constructing larger (counter)examples from smaller ones. The second motivation is to make progress towards a polynomial-time algorithm for deciding independence in the abovementioned maximal matroid. Nucleation-free graphs with implied nonedges, such as the families
The Experiment to Detect the Global Epoch of Reionization 21 cm Signal (EDGES) has reported evidence for an absorption feature in the sky-averaged radio background near 78 MHz. A cosmological interpretation of this signal corresponds to absorption of 21 cm photons by neutral hydrogen at $z \sim 17$. The large depth of the signal has been shown to require an excess radio background above the CMB and/or non-standard cooling processes in the IGM. Here, we explore the plausibility of a scenario in which the EDGES signal is back-lit by an excess radio background sourced from a population of radio-loud AGN at high redshift. These AGN could also explain the unexpected abundance of UV-bright objects observed at $z > 10$ by JWST. We find that producing enough radio photons to explain the EDGES depth requires that nearly all high-$z$ UV-bright objects down to $M_{\mathrm{UV}} \gtrsim -15$ are radio-loud AGN and that the UV density of such objects declines by at most 1.5 orders of magnitude between $z = $10 and 20. In addition, the fraction of X-ray photons escaping these objects must be $\lesssim 1\%$ of their expected intrinsic production rate to prevent the absorption signal being washe
3D-printing of radiofrequency (RF) cavity resonators could provide a cost-effective solution that enables rapid prototyping and design flexibility compared to traditional fabrication of full-metal cavities. In this work, the feasibility of fabrication of a useful multi-mode GHz cavity is explored. Two kinds of plastics, two slicing approaches and two metal coating techniques were used to build a series of clamped cavities with thin inner copper surface on otherwise 3D printed plastic surface. The cavities were then bench-tested to identify spatial field distributions, operating frequencies and quality factors (Q-factor). Pros and cons of the used fabrication approaches were identified and understood, and the performance of longitudinally sliced painted cavity design demonstrated considerable practicality of 3D-printing approach in designing rf systems.
Cyanopolyynes are among the largest and most commonly observed interstellar Complex Organic Molecules in star-forming regions. They are believed to form primarily in the gas-phase, but their formation routes are not well understood. We present a comprehensive study of the gas-phase formation network of cyanobutadiyne, HC$_5$N, based on new theoretical calculations, kinetics experiments, astronomical observations, and astrochemical modeling. We performed new quantum mechanics calculations for six neutral-neutral reactions in order to derive reliable rate coefficients and product branching fractions. We also present new CRESU data on the rate coefficients of three of these reactions (C$_3$N + C$_2$H$_2$, C$_2$H + HC$_3$N, CN + C$_4$H$_2$) obtained at temperatures as low as 24 K. In practice, six out of nine reactions currently used in astrochemical models have been updated in our reviewed network. We also report the tentative detection of the $^{13}$C isotopologues of HC$_5$N in the L1544 prestellar core. We derived a lower limit of $^{12}$C/$^{13}$C > 75 for the HC$_5$N isotopologues, which does not allow to bring new constraints to the HC$_5$N chemistry. Finally, we verified the
High-speed optical imaging of dynamic neuronal activity is essential yet challenging in neuroscience. While calcium imaging has been firmly established as a workhorse technique for monitoring neuronal activity, its limited temporal resolution and indirect measurement restrict its ability to capture rapid inhibitory and excitatory events and subthreshold voltage oscillations. In contrast, voltage imaging directly measures membrane potential fluctuations, providing a comprehensive and precise representation of neuronal circuit dynamics. Recent advancements in voltage-sensitive dyes and, particularly, genetically encoded voltage indicators have significantly enhanced the feasibility of voltage imaging, prompting the development of advanced fluorescence microscopy methods optimized for high-speed acquisition. However, achieving millisecond-scale temporal resolution remains challenging due to inherent trade-offs among imaging speed, spatial resolution, and signal-to-noise ratio. Conventional raster-scanning approaches, including confocal microscopy, are fundamentally limited by their slow frame rates, precluding the capture of rapid neuronal events from multiple neurons simultaneously.
We present the design and characterization of a dual-mode radiofrequency (rf) cavity, a novel electromagnetic structure with potential benefits such as compactness, efficiency, cost reduction and multifunctionality. The cavity was designed to balance the dual-mode structure considering several factors, such as mode frequencies, quality factor (Q-factor), and minimizing cross talk between couplers. We preformed various tests to verify that this cavity preformed as expected compared to simulated results. As exampled here, a combination of the the fundamental mode TM 010 and the TM 011 mode, tuned to a harmonic of the fundamental, was realized to linearize the off-crest electric field, thereby enabling concurrent bunching and acceleration of charged particle (e.g. electrons) beam in high power systems. The reduction in the number of cavities required to bunch and accelerate promises cost and space savings over conventional approaches. This research lays the foundation for further exploration of multi-mode cavity applications and optimization for specific use cases, with potential implications for a wide range of fields including quantum information platforms.
We consider étale Hausdorff groupoids in which the interior of the isotropy is abelian. We prove that the norms of the images under regular representations, of elements of the reduced groupoid $C^*$-algebra whose supports are contained in the interior of the isotropy vary upper semicontinuously. This corrects an error in [T.M. Carlsen, E. Ruiz, A. Sims and M. Tomforde, "Reconstruction of groupoids and C*-rigidity of dynamical systems," Adv. Math 390 (2021), 107923].
The age of Artificial Intelligence (AI) is marked by its transformative "generative" capabilities, distinguishing it from prior iterations. This burgeoning characteristic of AI has enabled it to produce new and original content, inherently showcasing its creative prowess. This shift challenges and requires a recalibration in the realm of arts education, urging a departure from established pedagogies centered on human-driven image creation. The paper meticulously addresses the integration of AI tools, with a spotlight on Stable Diffusion (SD), into university arts curricula. Drawing from practical insights gathered from workshops conducted in July 2023, which culminated in an exhibition of AI-driven artworks, the paper aims to provide a roadmap for seamlessly infusing these tools into academic settings. Given their recent emergence, the paper delves into a comprehensive overview of such tools, emphasizing the intricate dance between artists, developers, and researchers in the open-source AI art world. This discourse extends to the challenges and imperatives faced by educational institutions. It presents a compelling case for the swift adoption of these avant-garde tools, underscorin
We prove a conjectured graph theoretic characterization of a geometric property of 3 dimensional linkages posed 15 years ago by Sitharam and Gao, motivated by their equivalent characterization for $d\le 2$ that does not generalize to $d\ge 3$. A linkage $(G,\ell)$ contains a finite simple undirected graph $G$ and a map $\ell$ that assigns squared Euclidean lengths to the edges of $G$. A \emph{$d$-realization} of $(G,\ell)$ is an assignment of points in $\mathbb{R}^d$ to the vertices of $G$ for which pairwise squared distances between points agree with $\ell$. For any positive integer $d \leq 3$, we characterize pairs $(G,f)$, where $f$ is a nonedge of $G$, such that, for any linkage $(G,\ell)$, the lengths attained by $f$ form a single interval - over the (typically a disconnected set of) $d$-realizations of $(G,\ell)$. Although related to the minor closed class of $d$-flattenable graphs, the class of pairs $(G,f)$ with the above property is not closed under edge deletions, has no obvious well quasi-ordering, and there are infinitely many minimal graph-nonedge pairs - with respect to edge contractions - in the complement class. Our characterization overcomes these obstacles, is bas
Black holes are objects that have a large mass and curve space time, characterized by their event horizon and singularity. Recently, an interesting concept of analogous black holes has emerged in the field of condensed matter physics. In this work, the possibility of realizing analogous black holes in topological material is Ni$_3$In$_2$X$_2$ (X = S, Se) discussed. This work shows that the type-III Dirac cones of the material can lead to the emergence of an event horizon and the formation of a black hole-like region near the Dirac point. In addition, possible experimental signatures of such a system are discussed and the potential implications of an analogous black hole for the study of black hole physics in condensed matter systems.
The engineering of new states of matter through Floquet driving has revolutionized the field of condensed matter physics. This technique enables the creation of hybrid topological states and ordered phases that are absent in normal systems. Crystalline structures, exemplifying spatially ordered systems under periodic driving, have been extensively studied. However, the recent focus has shifted towards discrete time crystals (DTCs), periodically driven quantum many-body systems that break time translation symmetry under specific conditions. In this paper, we explore the theoretical predictions, experimental realizations, and emerging possibilities of utilizing DTCs on quantum computers. Additionally, the formation of time varying tesseracts using discrete time crystals is presented, allowing for the investigation of time translational symmetry in higher-dimensional lattice systems.
The study of topologically ordered states have given rise to a growing interest in symmetry protected states in quantum matter. Recently, this theory has been extended to quantum many body systems which demonstrate ordered states at low temperature. An example of this is the discrete time crystal (DTC) which has been demonstrated in a real quantum computer and in driven systems. These states are periodic in time and are protected to disorder to a certain extent. In general, DTC can be classified into two phases, the stable many body localization (MBL) state, and the disordered thermal state. This work demonstrates by generalizing DTC to 2 dimensions, there is an decrease in thermal noise and an increase in the operating range of the MBL range in the presence of disorder.
The fractional quantum hall effect (FQHE) is a milestone of modern day physics, its disovery paved the way for the study of fractional charges which do not obey abelian physics. However, all FQHE require an external magnetic field in order for there to be fractionally charged electrons. This work shows a theoretical study of strongly interaction electrons in the Kagome lattice with magnetism. Under optimal conditions, a band gap opens on the surface of the material which hosts fractional Fermions. These fractional Fermions form composite quasi-electrons without the need for external magnetic field. These states are predicted to host the FQHE without the need for an external magnetic field.
The PageRank algorithm is used to rank web pages by their importance. Since its development, the PageRank algorithm is a critical and fundamental part of search engines today. PageRank is a graph-based algorithm that ranks pages based on how many other pages link to them. This work develops a variational quantum version of the PageRank algorithm and compares the performance of the two algorithms. It is found that quantum PageRank performs better at ranking websites than the normal PageRank algorithm
Kagome Chern Magnets are lattices which host a Chern state when they are non-magnetic and a Chern Gap when they become magnetic. These Chern gaps can host fractional Chern Fermions. There has been extensive research in the RMn$_6$Sn$_6$ system which hosts a small Chern gap with many trivial bands surrounding them. RMn$_6$Ge$_6$ is studied theoretically in order to find a more insulating Kagome magnetic with fewer bands near the Chern gap.
Edge detection and image filters are commonly used in computer vision. However, they have never been applied to the data analysis of angle-resolved photoemission spectroscopy (ARPES) data before in a systematic fashion. In this paper we will use the Sobel, Laplacian of a gaussian (LoG), Canny, Prewitt, Roberts, and fuzzy logic methods for edge detection in the ARPES results of HfP2, ZrSiS, and Hf2Te2P2. We find that the Canny filter is the best method for edge detection of noisy data that is typical of ARPES measurements, while the other edge detection techniques are not able to correctly detect ARPES bands.
We present an open source software package SpectroLab a Matlab-based tool developed in 2018 for the analysis of spectroscopic data. In this package, there are tools for derivative analysis, stacked energy contours, stacked plots for theory, 3D volumetric plots, core level analysis, and derivative analysis. The package can currently be used for angle-resolved photoemission spectroscopy (ARPES) data, with the ability to also be used for other spectroscopic data in the future. We apply this program to the HfP2, ZrSiS, and Hf2Te2P systems to demonstrate its robustness.
This work proposes an algorithm for fMRI data analysis for the classification of ADHD disorders. There have been several breakthroughs in the analysis of fMRI via 3D convolutional neural networks (CNNs). With these new techniques it is possible to preserve the 3D spatial data of fMRI data. Additionally there have been recent advances in the use of 3D generative adversarial neural networks (GANs) for the generation of normal MRI data. This work utilizes multi modal 3D CNNs with data augmentation from 3D GAN for ADHD prediction from fMRI. By leveraging a 3D-GAN it would be possible to use deepfake data to enhance the accuracy of 3D CNN classification of brain disorders. A comparison will be made between a time distributed single modal 3D CNN model for classification and the modified multi modal model with MRI data as well.