Search

W Serpentis is an eclipsing binary system and the prototype of the Serpentid class of variable stars. These are interacting binaries experiencing intense mass transfer and mass loss. However, the identities and properties of both stars in W Ser remain a mystery. Here we present an observational analysis of high quality, visible-band spectroscopy made with the Apache Point Observatory 3.5 m telescope and ARCES spectrograph plus the first near-IR, long-baseline interferometric observations obtained with the CHARA Array. We present examples of the appearance and radial velocities of the main spectral components: prominent emission lines, strong shell absorption lines, and weak absorption lines. We show that some of the weak absorption features are associated with the cool mass donor, and we present the first radial velocity curve for the donor star. The donor's absorption lines are rotationally broadened, and we derive a ratio of donor to gainer mass of 0.36 +/- 0.09 based on the assumptions that the donor fills its Roche lobe and rotates synchronously with the orbit. We use a fit of the ASAS light curve to determine the orbital inclination and mass estimates of 2.0 and 5.7 solar masses for the donor and gainer, respectively. The partially resolved interferometric measurements of orbital motion are consistent with our derived orbital properties and the distance from Gaia EDR3. Spectroscopic evidence indicates that the gainer is enshrouded in an opaque disk that channels the mass transfer stream into an outflow through the L3 region and into a circumbinary disk., Comment: Published in ApJ

Many classical Be stars acquire their very rapid rotation by mass and angular-momentum transfer in massive binaries. Short-lived intermediate-phase objects have only been discovered recently. Data archives and the literature have been searched for additional candidates exhibiting this patterns. Thirteen candidates were identified at various confidence levels. Adding to the two known systems identified as classical Be star+pre-subdwarf binaries (LB-1 and HR6819), two more (V742Cas, HD44637) could be confirmed with interferometry, with V742Cas setting a new record for the smallest visually observed angular semi-major axis, at a=0.663mas. Two further ones (V447Sct, V1362Cyg) are not resolved interferometrically, but other evidence puts them at the same confidence level as LB-1. V2174Cyg is a candidate with very high confidence, but was not observed interferometrically. The remaining ones are either candidates with varying levels of confidence. Of a mostly magnitude complete sample of 328 Be stars, 0.5-1% are found to have recently completed the mass overflow that led to their formation. Another 5% are systems with compact subdwarf companions, i.e., further evolved after a previous overflow, and possibly two more percent harbor white dwarfs. All these systems are of early B-subtypes, however, and if the original sample is restricted to early subtypes (136 objects), these percentages increase by a factor of about 2.5, while dropping to zero for the mid and late subtypes (together 204 objects). This strongly suggests that early- vs. mid- and late-type Be stars have differently weighted channels to acquire their rapid rotation, namely binary interaction vs. evolutionary spin-up., Comment: Accepted for publication by A&A

Zero-shot capabilities of large language models make them powerful tools for solving a range of tasks without explicit training. It remains unclear, however, how these models achieve such performance, or why they can zero-shot some tasks but not others. In this paper, we shed some light on this phenomenon by defining and investigating algorithmic stability in language models -- changes in problem-solving strategy employed by the model as a result of changes in task specification. We focus on a task where algorithmic stability is needed for generalization: two-operand arithmetic. Surprisingly, we find that Gemma-2-2b employs substantially different computational models on closely related subtasks, i.e. four-digit versus eight-digit addition. Our findings suggest that algorithmic instability may be a contributing factor to language models' poor zero-shot performance across certain logical reasoning tasks, as they struggle to abstract different problem-solving strategies and smoothly transition between them., Comment: 10 pages, 5 figures

We present an application of our new theoretical formulation of quantum dynamics, moment propagation theory (MPT) (Boyer et al., J. Chem. Phys. 160, 064113 (2024)), for employing machine-learning techniques to simulate the quantum dynamics of electrons. In particular, we use real-time time-dependent density functional theory (RT-TDDFT) simulation in the gauge of the maximally localized Wannier functions (MLWFs) for training the MPT equation of motion. Spatially-localized time-dependent MLWFs provide a concise representation that is particularly convenient for the MPT expressed in terms of increasing orders of moments. The equation of motion for these moments can be integrated in time while the analytical expressions are quite involved. In this work, machine-learning techniques were used to train the the second-order time derivatives of the moments using first-principles data from the RT-TDDFT simulation, and this MPT enabled us to perform electron dynamics efficiently. The application to computing optical absorption spectrum for various systems was demonstrated as a proof-of-principles example of this approach. In addition to isolated molecules (water, benzene, and ethene), condensed matter systems (liquid water and crystalline silicon) were studied, and we also explored how the principle of the nearsightedness of electrons can be employed in this context.

A hallmark of intelligence is the ability to flesh out underspecified situations using "common sense." We propose to extract that common sense from large language models (LLMs), in a form that can feed into probabilistic inference. We focus our investigation on $\textit{guesstimation}$ questions such as "How much are Airbnb listings in Newark, NJ?" Formulating a sensible answer without access to data requires drawing on, and integrating, bits of common knowledge about how $\texttt{Price}$ and $\texttt{Location}$ may relate to other variables, such as $\texttt{Property Type}$. Our framework answers such a question by synthesizing an $\textit{ad hoc}$ probabilistic model. First we prompt an LLM to propose a set of random variables relevant to the question, followed by moment constraints on their joint distribution. We then optimize the joint distribution $p$ within a log-linear family to maximize the overall constraint satisfaction. Our experiments show that LLMs can successfully be prompted to propose reasonable variables, and while the proposed numerical constraints can be noisy, jointly optimizing for their satisfaction reconciles them. When evaluated on probabilistic questions derived from three real-world tabular datasets, we find that our framework performs comparably to a direct prompting baseline in terms of total variation distance from the dataset distribution, and is similarly robust to noise.

To track tumors during surgery, information from preoperative CT scans is used to determine their position. However, as the surgeon operates, the tumor may be deformed which presents a major hurdle for accurately resecting the tumor, and can lead to surgical inaccuracy, increased operation time, and excessive margins. This issue is particularly pronounced in robot-assisted partial nephrectomy (RAPN), where the kidney undergoes significant deformations during operation. Toward addressing this, we introduce a occupancy network-based method for the localization of tumors within kidney phantoms undergoing deformations at interactive speeds. We validate our method by introducing a 3D hydrogel kidney phantom embedded with exophytic and endophytic renal tumors. It closely mimics real tissue mechanics to simulate kidney deformation during in vivo surgery, providing excellent contrast and clear delineation of tumor margins to enable automatic threshold-based segmentation. Our findings indicate that the proposed method can localize tumors in moderately deforming kidneys with a margin of 6mm to 10mm, while providing essential volumetric 3D information at over 60Hz. This capability directly enables downstream tasks such as robotic resection., Comment: Accepted at IROS 2024

We introduce INQUIRE, a text-to-image retrieval benchmark designed to challenge multimodal vision-language models on expert-level queries. INQUIRE includes iNaturalist 2024 (iNat24), a new dataset of five million natural world images, along with 250 expert-level retrieval queries. These queries are paired with all relevant images comprehensively labeled within iNat24, comprising 33,000 total matches. Queries span categories such as species identification, context, behavior, and appearance, emphasizing tasks that require nuanced image understanding and domain expertise. Our benchmark evaluates two core retrieval tasks: (1) INQUIRE-Fullrank, a full dataset ranking task, and (2) INQUIRE-Rerank, a reranking task for refining top-100 retrievals. Detailed evaluation of a range of recent multimodal models demonstrates that INQUIRE poses a significant challenge, with the best models failing to achieve an mAP@50 above 50%. In addition, we show that reranking with more powerful multimodal models can enhance retrieval performance, yet there remains a significant margin for improvement. By focusing on scientifically-motivated ecological challenges, INQUIRE aims to bridge the gap between AI capabilities and the needs of real-world scientific inquiry, encouraging the development of retrieval systems that can assist with accelerating ecological and biodiversity research. Our dataset and code are available at https://inquire-benchmark.github.io, Comment: Published in NeurIPS 2024, Datasets and Benchmarks Track

Species range maps (SRMs) are essential tools for research and policy-making in ecology, conservation, and environmental management. However, traditional SRMs rely on the availability of environmental covariates and high-quality species location observation data, both of which can be challenging to obtain due to geographic inaccessibility and resource constraints. We propose a novel approach combining millions of citizen science species observations with textual descriptions from Wikipedia, covering habitat preferences and range descriptions for tens of thousands of species. Our framework maps locations, species, and text descriptions into a common space, facilitating the learning of rich spatial covariates at a global scale and enabling zero-shot range estimation from textual descriptions. Evaluated on held-out species, our zero-shot SRMs significantly outperform baselines and match the performance of SRMs obtained using tens of observations. Our approach also acts as a strong prior when combined with observational data, resulting in more accurate range estimation with less data. We present extensive quantitative and qualitative analyses of the learned representations in the context of range estimation and other spatial tasks, demonstrating the effectiveness of our approach., Comment: NeurIPS 2024

The abundance and distribution of metal in asteroid surfaces can be constrained from thermal emission measurements at radio wavelengths, informing our understanding of planetesimal differentiation processes. We observed the M-type asteroid (22) Kalliope and its moon Linus in thermal emission at 1.3, 9, and 20 mm with the Atacama Large Millimeter/submillimeter Array (ALMA) and the Karl G. Jansky Very Large Array (VLA) over most of Kalliope's rotation period. The 1.3 mm data provide ~30 km resolution on the surface of Kalliope, while both the 1.3 and 9 mm data resolve Linus from Kalliope. We find a thermal inertia for Kalliope of 116$^{+326}_{-91}$ J m$^{-2}$ s$^{-0.5}$ K$^{-1}$ and emissivities of 0.65$\pm$0.02 at 1.3 mm, 0.56$\pm$0.03 at 9 mm, and 0.77$\pm$0.02 at 20 mm. Kalliope's millimeter wavelength emission is suppressed compared to its centimeter wavelength emission, and is also depolarized. We measure emissivities for Linus of 0.73$\pm$0.04 and 0.85$\pm$0.17 at 1.3 and 9 mm respectively, indicating a less metal-rich surface composition for Linus. Spatial variability in Kalliope's emissivity reveals a region in the northern hemisphere with a high dielectric constant, suggestive of enhanced metal content. These results are together consistent with a scenario in which Linus formed from reaggregated ejecta from an impact onto a differentiated Kalliope, leaving Kalliope with a higher surface metal content than Linus, which is distributed heterogeneously across its surface. The low emissivity and lack of polarization suggest a reduced regolith composition where iron is in the form of metallic grains and constitutes ~25% of the surface composition., Comment: Accepted to PSJ

Stars with initial masses larger than 8 solar masses undergo substantial mass loss through mechanisms that remain elusive. Unraveling the origins of this mass loss is important for comprehending the evolutionary path of these stars, the type of supernova explosion and whether they become neutron stars or black hole remnants. In 2022 December, RW Cep experienced the Great Dimming in its visible brightness, presenting a unique opportunity to understand mass loss mechanisms. Our previous observations of RW Cep from the CHARA Array, taken during the dimming phase, show a compelling asymmetry in the star images, with a darker zone on the west side of the star indicating presence of dust in front of the star in our line of sight. Here, we present multi-epoch observations from CHARA while the star re-brightened in 2023. We created images using three image reconstruction methods and an analytical model fit. Comparisons of images acquired during the dimming and re-brightening phases reveal remarkable differences. Specifically, the west side of RW Cep, initially obscured during the dimming phase, reappeared during the subsequent re-brightening phase and the measured angular diameter became larger by 8%. We also observed image changes from epoch to epoch while the star is brightening indicating the time evolution of dust in front of the star. We suggest that the dimming of RW Cep was a result from a recent surface mass ejection event, generating a dust cloud that partially obstructed the stellar photosphere., Comment: Accepted by The Astrophysical Journal Letters (ApJL)

Deep learning based approaches are now widely used across biophysics to help automate a variety of tasks including image segmentation, feature selection, and deconvolution. However, the presence of multiple competing deep learning architectures, each with its own unique advantages and disadvantages, makes it challenging to select an architecture best suited for a specific application. As such, we present a comprehensive comparison of common models. Here, we focus on the task of segmentation assuming the typically small training dataset sizes available from biophysics experiments and compare the following four commonly used architectures: convolutional neural networks, U-Nets, vision transformers, and vision state space models. In doing so, we establish criteria for determining optimal conditions under which each model excels, thereby offering practical guidelines for researchers and practitioners in the field.

TICAS partnered with higher education researcher Dr. Donald E. Heller to examine the "affordability gap" that students are facing when paying for college. The report uses federal data to determine the so-called "college affordability gap" in three states--California, Michigan, and New York--with a focus on students who are eligible for the federal Pell Grant. The affordability gap is defined in this analysis as the gap between a student's total costs of attending college--including tuition, fees, housing, food, and other expenses--and the amount of non-loan aid available to them: a combination of family resources, grant aid, and earnings from a reasonable number of work hours. The report then examines how much the example states would need to invest close this gap for Pell-eligible students, with the goal of a $0 gap for such students. This analysis builds on TICAS' ongoing efforts to build a debt-free future for all students.

Modern artificial intelligence (AI) systems are powered by foundation models. This paper presents a new set of foundation models, called Llama 3. It is a herd of language models that natively support multilinguality, coding, reasoning, and tool usage. Our largest model is a dense Transformer with 405B parameters and a context window of up to 128K tokens. This paper presents an extensive empirical evaluation of Llama 3. We find that Llama 3 delivers comparable quality to leading language models such as GPT-4 on a plethora of tasks. We publicly release Llama 3, including pre-trained and post-trained versions of the 405B parameter language model and our Llama Guard 3 model for input and output safety. The paper also presents the results of experiments in which we integrate image, video, and speech capabilities into Llama 3 via a compositional approach. We observe this approach performs competitively with the state-of-the-art on image, video, and speech recognition tasks. The resulting models are not yet being broadly released as they are still under development.

Dynamic critical fluctuations in magnetic materials encode important information about magnetic ordering in the associated critical exponents. Using nitrogen-vacancy centers in diamond, we implement $T_2$ (spin-decoherence) noise magnetometry to study critical dynamics in a 2D Van der Waals magnet CrSBr. By analyzing NV decoherence on time scales approaching the characteristic correlation time $\tau_c$ of critical fluctuations, we extract the critical exponent $\nu$ for the correlation length. Our result deviates from the Ising prediction and highlights the role of long-range dipolar interactions in 2D CrSBr. Furthermore, analyzing the divergence of the correlation length suggests the possibility of 2D-XY criticality in CrSBr in a temperature window near $T_C$ where static magnetic domains are absent. Our work provides a first demonstration of $T_2$ noise magnetometry to quantitatively analyze critical scaling behavior in 2D materials., Comment: 19 pages main text, 4 main text figures, 26 pages Supplementary Materials, 13 Supplementary figures

One of the keys to the success of collaborative learning is balanced participation by all learners, but this does not always happen naturally. Pedagogical robots have the potential to facilitate balance. However, it remains unclear what participation balance robots should aim at; various metrics have been proposed, but it is still an open question whether we should balance human participation in human-human interactions (HHI) or human-robot interactions (HRI) and whether we should consider robots' participation in collaborative learning involving multiple humans and a robot. This paper examines collaborative learning between a pair of students and a teachable robot that acts as a peer tutee to answer the aforementioned question. Through an exploratory study, we hypothesize which balance metrics in the literature and which portions of dialogues (including vs. excluding robots' participation and human participation in HHI vs. HRI) will better predict learning as a group. We test the hypotheses with another study and replicate them with automatically obtained units of participation to simulate the information available to robots when they adaptively fix imbalances in real-time. Finally, we discuss recommendations on which metrics learning science researchers should choose when trying to understand how to facilitate collaboration., Comment: To appear in Seventeenth International Conference on Educational Data Mining (EDM 2024)

Main-belt asteroid (16) Psyche is the largest M-type asteroid, a class of object classically thought to be the metal cores of differentiated planetesimals and the parent bodies of the iron meteorites. de Kleer, Cambioni, and Shepard (2021, https://doi.org/10.3847/psj/ac01ec) presented new data from the Atacama Large Millimeter Array (ALMA), from which they derived a global best-fit thermal inertia and dielectric constant for Psyche, proxies for regolith particle size, porosity, and/or metal content, and observed thermal anomalies that could not be explained by surface albedo variations only. Motivated by this, here we fit a model to the same ALMA data set that allows dielectric constant and thermal inertia to vary across the surface. We find that Psyche has a heterogeneous surface in both dielectric constant and thermal inertia but, intriguingly, we do not observe a direct correlation between these two properties over the surface. We explain the heterogeneity in dielectric constant as being due to variations in the relative abundance of metal and silicates. Furthermore, we observe that the lowlands of a large depression in Psyche's shape have distinctly lower thermal inertia than the surrounding highlands. We propose that the latter could be explained by a thin mantle of fine regolith, fractured bedrock, and/or implanted silicate-rich materials covering an otherwise metal-rich surface. All these scenarios are indicative of a collisionally evolved world., Comment: Paper is published open-access here: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2021JE007091

Prostate cancer diagnosis continues to encounter challenges, often due to imprecise needle placement in standard biopsies. Several control strategies have been developed to compensate for needle tip prediction inaccuracies, however none were compared against each other, and it is unclear whether any of them can be safely and universally applied in clinical settings. This paper compares the performance of two resolved-rate controllers, derived from a mechanics-based and a data-driven approach, for bevel-tip needle control using needle shape manipulation through a template. We demonstrate for a simulated 12-core biopsy procedure under model parameter uncertainty that the mechanics-based controller can better reach desired targets when only the final goal configuration is presented even with uncertainty on model parameters estimation, and that providing a feasible needle path is crucial in ensuring safe surgical outcomes when either controller is used for needle shape manipulation.

We present a novel theoretical formulation for performing quantum dynamics in terms of moments within the single-particle description. By expressing the quantum dynamics in terms of increasing orders of moments, instead of single-particle wave functions as generally done in time-dependent density functional theory, we describe an approach for reducing the high computational cost of simulating the quantum dynamics. The equation of motion is given for the moments by deriving analytical expressions for the first-order and second-order time derivatives of the moments, and a numerical scheme is developed for performing quantum dynamics by expanding the moments in the Taylor series as done in classical molecular dynamics simulation. We propose a few numerical approaches using this theoretical formalism on a simple one-dimensional model system, for which an analytically exact solution can be derived. Application of the approaches to an anharmonic system is also discussed to illustrate their generality. We also discuss the use of an artificial neural network model to circumvent the numerical evaluation of the second-order time derivatives of the moments, as analogously done in the context of classical molecular dynamics simulations.

Understanding how the electronic excitation of DNA changes in response to different high-energy particles is central to advancing ion beam cancer therapy. While protons have been the predominant ion of choice in beam cancer therapy, heavier ions, particularly carbon ions, have drawn significant attention over the past decade. Carbon ions are expected to transfer larger amounts of energy accordingly to the linear response theory. However, the molecular-level details of the electronic excitation under heavier ion irradiation remain unknown. In this work, we use real-time time-dependent density functional theory simulations to examine the quantum-mechanical details of DNA electronic excitations in water under proton, alpha-particle, and carbon ion irradiation. Our results show that the energy transfer does indeed increase for the heavier ions, while the excitation remains highly conformal. However, the increase in the energy transfer rate, measured by electronic stopping power, does not match the prediction by the linear response model, even when accounting for the velocity-dependence of the irradiating ions charge. The simulations also reveal that while the number of holes generated on DNA increases for heavier ions, the increase is only partially responsible for the larger stopping power. Larger numbers of highly energetic holes formed from the heavier ions also contribute significantly to the increased electronic stopping power., Comment: 13 pages

Although the relationship between temperature and emotional states has been investigated in the field of haptics, it remains unknown if, or in what direction, temperature affects emotional states. We approach this question at the intersection of haptics and psychology using a custom-built thermal device and emotional responses based on photos from the International Affective Picture System (IAPS) library. Unlike past research, this study incorporates deception and a control (i.e., neutral temperature) condition. One hundred and twenty naive subjects reported their emotional responses to fifty-six images varying on normative arousal and valence ratings while being exposed to a cool~(30{\deg}C), neutral (33{\deg}C), or warm (36{\deg}C) temperature applied to the upper back. Participants exposed to warm temperatures reported higher arousal ratings in some image categories than participants exposed to neutral or cool temperatures. Valence ratings were decreased in warm conditions compared to neutral conditions. The emotion wheel was used as a complementary method of affective response measurement, and exploratory analysis methods were implemented. Although the valence and arousal showed statistical significance, the emotion wheel results did not demonstrate any significant differences between the temperature conditions.