Search

Recent advancements in large language models (LLMs) have shown potential for transforming data processing in healthcare, particularly in understanding complex clinical narratives. This study evaluates the efficacy of zero-shot LLMs in summarizing long clinical texts that require temporal reasoning, a critical aspect for comprehensively capturing patient histories and treatment trajectories. We applied a series of advanced zero-shot LLMs to extensive clinical documents, assessing their ability to integrate and accurately reflect temporal dynamics without prior task-specific training. While the models efficiently identified key temporal events, they struggled with chronological coherence over prolonged narratives. The evaluation, combining quantitative and qualitative methods, highlights the strengths and limitations of zero-shot LLMs in clinical text summarization. The results suggest that while promising, zero-shot LLMs require further refinement to effectively support clinical decision-making processes, underscoring the need for enhanced model training approaches that better capture the nuances of temporal information in long context medical documents.

We present nucleosynthesis and light-curve predictions for a new site of the rapid neutron capture process ($r$-process) from magnetar giant flares (GFs). Motivated by observations indicating baryon ejecta from GFs, Cehula et al. (2024) proposed mass ejection occurs after a shock is driven into the magnetar crust during the GF. We confirm using nuclear reaction network calculations that these ejecta synthesize moderate yields of third-peak $r$-process nuclei and more substantial yields of lighter $r$-nuclei, while leaving a sizable abundance of free neutrons in the outermost fastest expanding ejecta layers. The final $r$-process mass fraction and distribution are sensitive to the relative efficiencies of $\alpha$-capture and $n$-capture freeze-outs. We use our nucleosynthesis output in a semi-analytic model to predict the light curves of novae breves, the transients following GFs powered by radioactive decay. For a baryonic ejecta mass similar to that inferred of the 2004 Galactic GF from SGR 1806-20, we predict a peak UV/optical luminosity of $\sim 10^{39}$-$10^{40}\,\rm erg\,s^{-1}$ at $\sim 10$-$15$ minutes, rendering such events potentially detectable following a gamma-ray trigger by wide-field transient monitors such as ULTRASAT/UVEX to several Mpc. The peak luminosity and timescale of the transient increase with the GF strength due to the larger ejecta mass. Although GFs likely contribute 1-10% of the total Galactic $r$-process budget, their short delay-times relative to star-formation make them an attractive source to enrich the earliest generations of stars.

CairoZero is a programming language for running decentralized applications (dApps) at scale. Programs written in the CairoZero language are compiled to machine code for the Cairo CPU architecture and cryptographic protocols are used to verify the results of execution efficiently on blockchain. We explain how we have extended the CairoZero compiler with tooling that enables users to prove, in the Lean 3 proof assistant, that compiled code satisfies high-level functional specifications. We demonstrate the success of our approach by verifying primitives for computation with the secp256k1 and secp256r1 curves over a large finite field as well as the validation of cryptographic signatures using the former. We also verify a mechanism for simulating a read-write dictionary data structure in a read-only setting. Finally, we reflect on our methodology and discuss some of the benefits of our approach.

We define a cap in the affine geometry AG(n,2) to be a subset in which every collection of four points is in general position. In this paper, we classify, up to affine equivalence, all caps in AG(7,2) of size k greater than or equal to 10. In particular, we show that there are two equivalence classes of 10-caps and one equivalence class of 11-caps, none of which are complete, and one equivalence class of 12-caps, which are both complete and of maximum size., Comment: 33 pages, 2 figures, 3 tables

In the quest to improve services, companies offer customers the option to interact with agents via texting. Such contact centers face unique challenges compared to traditional call centers, as measuring customer experience proxies like abandonment and patience involves uncertainty. A key source of this uncertainty is silent abandonment, where customers leave without notifying the system, wasting agent time and leaving their status unclear. Silent abandonment also obscures whether a customer was served or left. Our goals are to measure the magnitude of silent abandonment and mitigate its effects. Classification models show that 3%-70% of customers across 17 companies abandon silently. In one study, 71.3% of abandoning customers did so silently, reducing agent efficiency by 3.2% and system capacity by 15.3%, incurring $5,457 in annual costs per agent. We develop an expectation-maximization (EM) algorithm to estimate customer patience under uncertainty and identify influencing covariates. We find that companies should use classification models to estimate abandonment scope and our EM algorithm to assess patience. We suggest strategies to operationally mitigate the impact of silent abandonment by predicting suspected silent-abandonment behavior or changing service design. Specifically, we show that while allowing customers to write while waiting in the queue creates a missing data challenge, it also significantly increases patience and reduces service time, leading to reduced abandonment and lower staffing requirements., Comment: 75% of the paper is an updated version of arXiv:2304.11754

The origin of heavy elements synthesized through the rapid neutron capture process ($r$-process) has been an enduring mystery for over half a century. Cehula et al. (2024) recently showed that magnetar giant flares, among the brightest transients ever observed, can shock-heat and eject neutron star crustal material at high velocity, achieving the requisite conditions for an $r$-process. Patel et al. (in prep.) confirmed an $r$-process in these ejecta using detailed nucleosynthesis calculations. Radioactive decay of the freshly synthesized nuclei releases a forest of gamma-ray lines, Doppler broadened by the high ejecta velocities $v \gtrsim 0.1c$ into a quasi-continuous spectrum peaking around 1 MeV. Here, we show that the predicted emission properties (light-curve, fluence, and spectrum) match a previously unexplained hard gamma-ray signal seen in the aftermath of the famous December 2004 giant flare from the magnetar SGR 1806-20. This MeV emission component, rising to peak around 10 minutes after the initial spike before decaying away over the next few hours, is direct observational evidence for the synthesis of $\sim 10^{-6}M_{\odot}$ of $r$-process elements. The discovery of magnetar giant flares as confirmed $r$-process sites, contributing at least $\sim 1$-$10\%$ of the total Galactic abundances, has implications for the Galactic chemical evolution, especially at the earliest epochs probed by low-metallicity stars. It also implicates magnetars as potentially dominant sources of heavy cosmic rays. Characterization of the $r$-process emission from giant flares by resolving decay line features offers a compelling science case for NASA's forthcoming COSI nuclear spectrometer, as well as next-generation MeV telescope missions., Comment: 15 pages, 6 figures, 1 table

Quantum optics bridges esoteric notions of entanglement and superposition with practical applications like metrology and communication. Throughout, there is an interplay between information theoretic concepts such as entropy and physical considerations such as quantum system design, noise, and loss. Therefore, a fundamental result at the heart of these fields has numerous ramifications in development of applications and advancing our understanding of quantum physics. Our recent proof for the entanglement properties of states interfering with the vacuum on a beam splitter led to monotonicity and convexity properties for quantum states undergoing photon loss [Lupu-Gladstein et al., arXiv:2411.03423 (2024)] by breathing life into a decades-old conjecture. In this work, we extend these fundamental properties to measures of similarity between states, provide inequalities for creation and annihilation operators beyond the Cauchy-Schwarz inequality, prove a conjecture [Hertz et al., PRA 110, 012408 (2024)] dictating that nonclassicality through the quadrature coherence scale is uncertifiable beyond a loss of 50%, and place constraints on quasiprobability distributions of all physical states. These ideas can now circulate afresh throughout quantum optics., Comment: 12+4 pages, companion to arXiv:2411.03423, comments warmly welcomed

This research explores the beta decay of the proton-rich nucleus 23Al. The nucleus was generated at the National Superconducting Cyclotron Laboratory (NSCL) through projectile fragmentation, utilizing a primary beam of 36Ar ions directed at a 9Be target. Simultaneous measurements of proton emission and gamma rays were conducted using the GADGET detection system. The decay paths were carefully analyzed through beta-gamma , proton-gamma , and gamma -gamma coincidences, leading to the construction of a complete decay scheme for 23Al. The absolute beta branching ratios were determined, and log-ft values were calculated for transitions to 23Mg states. Additionally, proton branching ratios and the most precise half-life measurement of 23Al to date were obtained. The findings include the identification of 19 new gamma rays and the discovery of a new beta-delayed proton transition populating the third excited state of 22Na., Comment: Master's thesis 60 pages, 28 figures. Under the supervision of Dr. Moshe Friedman

We propose a decentralized market model in which agents can negotiate bilateral contracts. This builds on a similar, but centralized, model of trading networks introduced by Hatfield et al. in 2013. Prior work has established that fully-substitutable preferences guarantee the existence of competitive equilibria which can be centrally computed. Our motivation comes from the fact that prices in markets such as over-the-counter markets and used car markets arise from decentralized negotiation among agents, which has left open an important question as to whether equilibrium prices can emerge from agent-to-agent bilateral negotiations. We design a best response dynamic intended to capture such negotiations between market participants. We assume fully substitutable preferences for market participants. In this setting, we provide proofs of convergence for sparse markets (covering many real world markets of interest), and experimental results for more general cases, demonstrating that prices indeed reach equilibrium, quickly, via bilateral negotiations. Our best response dynamic, and its convergence behavior, forms an important first step in understanding how decentralized markets reach, and retain, equilibrium., Comment: Extended version of paper accepted at AAAI'25

How are AI assistants being used in the real world? While model providers in theory have a window into this impact via their users' data, both privacy concerns and practical challenges have made analyzing this data difficult. To address these issues, we present Clio (Claude insights and observations), a privacy-preserving platform that uses AI assistants themselves to analyze and surface aggregated usage patterns across millions of conversations, without the need for human reviewers to read raw conversations. We validate this can be done with a high degree of accuracy and privacy by conducting extensive evaluations. We demonstrate Clio's usefulness in two broad ways. First, we share insights about how models are being used in the real world from one million Claude.ai Free and Pro conversations, ranging from providing advice on hairstyles to providing guidance on Git operations and concepts. We also identify the most common high-level use cases on Claude.ai (coding, writing, and research tasks) as well as patterns that differ across languages (e.g., conversations in Japanese discuss elder care and aging populations at higher-than-typical rates). Second, we use Clio to make our systems safer by identifying coordinated attempts to abuse our systems, monitoring for unknown unknowns during critical periods like launches of new capabilities or major world events, and improving our existing monitoring systems. We also discuss the limitations of our approach, as well as risks and ethical concerns. By enabling analysis of real-world AI usage, Clio provides a scalable platform for empirically grounded AI safety and governance.

Two substantial technological advances have reshaped the public square in recent decades: first with the advent of the internet and second with the recent introduction of large language models (LLMs). LLMs offer opportunities for a paradigm shift towards more decentralized, participatory online spaces that can be used to facilitate deliberative dialogues at scale, but also create risks of exacerbating societal schisms. Here, we explore four applications of LLMs to improve digital public squares: collective dialogue systems, bridging systems, community moderation, and proof-of-humanity systems. Building on the input from over 70 civil society experts and technologists, we argue that LLMs both afford promising opportunities to shift the paradigm for conversations at scale and pose distinct risks for digital public squares. We lay out an agenda for future research and investments in AI that will strengthen digital public squares and safeguard against potential misuses of AI., Comment: 40 pages, 5 figures

We characterize sums of normal ultrafilters after the Magidor iteration of Prikry forcings over a discrete set of measurable cardinals. We apply this to show that the weak Ultrapower Axiom is not equivalent to the Ultrapower Axiom. We also construct a non-rigid ultrapower and two uniform ultrafilters on different cardinals that have the same ultrapower.

AI assistants are being created to help software engineers conduct a variety of coding-related tasks, such as writing, documenting, and testing code. We describe the use of the watsonx Code Assistant (WCA), an LLM-powered coding assistant deployed internally within IBM. Through surveys of two user cohorts (N=669) and unmoderated usability testing (N=15), we examined developers' experiences with WCA and its impact on their productivity. We learned about their motivations for using (or not using) WCA, we examined their expectations of its speed and quality, and we identified new considerations regarding ownership of and responsibility for generated code. Our case study characterizes the impact of an LLM-powered assistant on developers' perceptions of productivity and it shows that although such tools do often provide net productivity increases, these benefits may not always be experienced by all users., Comment: 21 pages, 3 figures. To be published in CHI EA 2025

Renewable and biodegradable plastics derived from soy protein isolate (SPI) offer a promising alternative to conventional petroleum-based plastics, particularly for film-grade bioplastics applications such as plastic bags. However, even with reinforcement from cellulose nanocrystals (CNCs), their mechanical properties including stiffness lag behind those of petroleum-based plastics. To identify pathways for improving CNC-reinforced SPI composites, we studied stiffening mechanisms by interpreting experimental data using homogenization models that accounted for CNC agglomeration and the formation of CNC/SPI interphases. To model effects of surface modification of CNCs with polydopamine (polyDOPA), we incorporated two key mechanisms: enhanced CNC dispersion and modified CNC-SPI interfacial interactions. Models accounted for interphases surrounding CNCs, arising from physicochemical interactions with the polyDOPA-modified CNC surfaces. Consistent wih experimental observations of polyDOPA modification enhancing mechanical properties through both increased spatial distribution of CNCs and matrix-filler interactions, results demonstrated that improved dispersion and interfacial bonding contribute to increased composite stiffness. Results highlight the potential of biodegradable CNC/SPI bio-nanocomposites as sustainable plastic alternatives, and suggest pathways for further enhancing their mechanical properties., Comment: 28 pages, 8 figures

In few-shot relation classification (FSRC), models must generalize to novel relations with only a few labeled examples. While much of the recent progress in NLP has focused on scaling data size, we argue that diversity in relation types is more crucial for FSRC performance. In this work, we demonstrate that training on a diverse set of relations significantly enhances a model's ability to generalize to unseen relations, even when the overall dataset size remains fixed. We introduce REBEL-FS, a new FSRC benchmark that incorporates an order of magnitude more relation types than existing datasets. Through systematic experiments, we show that increasing the diversity of relation types in the training data leads to consistent gains in performance across various few-shot learning scenarios, including high-negative settings. Our findings challenge the common assumption that more data alone leads to better performance and suggest that targeted data curation focused on diversity can substantially reduce the need for large-scale datasets in FSRC.

Cloud robotics enables robots to offload complex computational tasks to cloud servers for performance and ease of management. However, cloud compute can be costly, cloud services can suffer occasional downtime, and connectivity between the robot and cloud can be prone to variations in network Quality-of-Service (QoS). We present FogROS2-FT (Fault Tolerant) to mitigate these issues by introducing a multi-cloud extension that automatically replicates independent stateless robotic services, routes requests to these replicas, and directs the first response back. With replication, robots can still benefit from cloud computations even when a cloud service provider is down or there is low QoS. Additionally, many cloud computing providers offer low-cost spot computing instances that may shutdown unpredictably. Normally, these low-cost instances would be inappropriate for cloud robotics, but the fault tolerance nature of FogROS2-FT allows them to be used reliably. We demonstrate FogROS2-FT fault tolerance capabilities in 3 cloud-robotics scenarios in simulation (visual object detection, semantic segmentation, motion planning) and 1 physical robot experiment (scan-pick-and-place). Running on the same hardware specification, FogROS2-FT achieves motion planning with up to 2.2x cost reduction and up to a 5.53x reduction on 99 Percentile (P99) long-tail latency. FogROS2-FT reduces the P99 long-tail latency of object detection and semantic segmentation by 2.0x and 2.1x, respectively, under network slowdown and resource contention., Comment: IEEE/RSJ International Conference on Intelligent Robots and Systems 2024 Best Paper Finalist

Generative AI systems have shown impressive capabilities in creating text, code, and images. Inspired by the rich history of research in industrial ''Design for Assembly'', we introduce a novel problem: Generative Design-for-Robot-Assembly (GDfRA). The task is to generate an assembly based on a natural language prompt (e.g., ''giraffe'') and an image of available physical components, such as 3D-printed blocks. The output is an assembly, a spatial arrangement of these components, and instructions for a robot to build this assembly. The output must 1) resemble the requested object and 2) be reliably assembled by a 6 DoF robot arm with a suction gripper. We then present Blox-Net, a GDfRA system that combines generative vision language models with well-established methods in computer vision, simulation, perturbation analysis, motion planning, and physical robot experimentation to solve a class of GDfRA problems with minimal human supervision. Blox-Net achieved a Top-1 accuracy of 63.5% in the ''recognizability'' of its designed assemblies (eg, resembling giraffe as judged by a VLM). These designs, after automated perturbation redesign, were reliably assembled by a robot, achieving near-perfect success across 10 consecutive assembly iterations with human intervention only during reset prior to assembly. Surprisingly, this entire design process from textual word (''giraffe'') to reliable physical assembly is performed with zero human intervention., Comment: 8 pages, 7 Figures

Machine learning based surrogate models offer researchers powerful tools for accelerating simulation-based workflows. However, as standard datasets in this space often cover small classes of physical behavior, it can be difficult to evaluate the efficacy of new approaches. To address this gap, we introduce the Well: a large-scale collection of datasets containing numerical simulations of a wide variety of spatiotemporal physical systems. The Well draws from domain experts and numerical software developers to provide 15TB of data across 16 datasets covering diverse domains such as biological systems, fluid dynamics, acoustic scattering, as well as magneto-hydrodynamic simulations of extra-galactic fluids or supernova explosions. These datasets can be used individually or as part of a broader benchmark suite. To facilitate usage of the Well, we provide a unified PyTorch interface for training and evaluating models. We demonstrate the function of this library by introducing example baselines that highlight the new challenges posed by the complex dynamics of the Well. The code and data is available at https://github.com/PolymathicAI/the_well., Comment: 38th Conference on Neural Information Processing Systems (NeurIPS 2024) Track on Datasets and Benchmarks

Despite the significant strides made by generative AI in just a few short years, its future progress is constrained by the challenge of building modular and robust systems. This capability has been a cornerstone of past technological revolutions, which relied on combining components to create increasingly sophisticated and reliable systems. Cars, airplanes, computers, and software consist of components-such as engines, wheels, CPUs, and libraries-that can be assembled, debugged, and replaced. A key tool for building such reliable and modular systems is specification: the precise description of the expected behavior, inputs, and outputs of each component. However, the generality of LLMs and the inherent ambiguity of natural language make defining specifications for LLM-based components (e.g., agents) both a challenging and urgent problem. In this paper, we discuss the progress the field has made so far-through advances like structured outputs, process supervision, and test-time compute-and outline several future directions for research to enable the development of modular and reliable LLM-based systems through improved specifications.

This paper describes the robot technology behind an original performance that pairs a human dancer (Cuan) with an industrial robot arm for an eight-hour dance that unfolds over the timespan of an American workday. To control the robot arm, we combine a range of sinusoidal motions with varying amplitude, frequency and offset at each joint to evoke human motions common in physical labor such as stirring, digging, and stacking. More motions were developed using deep learning techniques for video-based human-pose tracking and extraction. We combine these pre-recorded motions with improvised robot motions created live by putting the robot into teach-mode and triggering force sensing from the robot joints onstage. All motions are combined with commercial and original music using a custom suite of python software with AppleScript, Keynote, and Zoom to facilitate on-stage communication with the dancer. The resulting performance contrasts the expressivity of the human body with the precision of robot machinery. Video, code and data are available on the project website: https://sites.google.com/playing.studio/breathless, Comment: 15 pages, 9 figures, accepted for ISRR (International Symposium of Robotics Research) 2024

Stripped-envelope supernovae (SESNe) represent a significant fraction of core-collapse supernovae, arising from massive stars that have shed their hydrogen and, in some cases, helium envelopes. The origins and explosion mechanisms of SESNe remain a topic of active investigation. In this work, we employ radiative-transfer simulations to model the light curves and spectra of a set of explosions of single, solar-metallicity, massive Wolf-Rayet (WR) stars with ejecta masses ranging from 4 to 11 Msun, that were computed from a turbulence-aided and neutrino-driven explosion mechanism. We analyze these synthetic observables to explore the impact of varying ejecta mass and helium content on observable features. We find that the light curve shape of these progenitors with high ejecta masses is consistent with observed SESNe with broad light curves but not the peak luminosities. The commonly used analytic formula based on rising bolometric light curves overestimates the ejecta mass of these high-initial-mass progenitor explosions by a factor up to 2.6. In contrast, the calibrated method by Haynie et al., which relies on late-time decay tails, reduces uncertainties to an average of 20% within the calibrated ejecta mass range.Spectroscopically, the He I 1.083 um line remains prominent even in models with as little as 0.02 Msun of helium. However, the strength of the optical He I lines is not directly proportional to the helium mass but instead depends on a complex interplay of factors such as 56Ni distribution, composition, and radiation field. Thus, producing realistic helium features requires detailed radiative transfer simulations for each new hydrodynamic model., Comment: 17 pages, 8 figures, accepted to APJ

It is typically assumed that for the successful use of machine learning algorithms, these algorithms should have a higher accuracy than a human expert. Moreover, if the average accuracy of ML algorithms is lower than that of a human expert, such algorithms should not be considered and are counter-productive. However, this is not always true. We provide strong statistical evidence that shows that even if a human expert is more accurate than a machine, an interaction with such a machine is beneficial when communication with the machine is non-public. The existence of a conflict between the user and ML model, and the private nature of user-AI communication will have the effect of making the user think about their decision and hence increase overall accuracy.

In the first part of the manuscript, we establish several consistency results concerning Woodin's $\HOD$ hypothesis and large cardinals around the level of extendibility. First, we prove that the first extendible cardinal can be the first strongly compact in HOD. We extend a former result of Woodin by showing that under the HOD hypothesis the first extendible cardinal is $C^{(1)}$-supercompact in HOD. We also show that the first cardinal-correct extendible may not be extendible, thus answering a question by Gitman and Osinski \cite[\S9]{GitOsi}. In the second part of the manuscript, we discuss the extent to which weak covering can fail below the first supercompact cardinal $\delta$ in a context where the HOD hypothesis holds. Answering a question of Cummings et al. \cite{CumFriGol}, we show that under the $\HOD$ hypothesis there are many singulars $\kappa<\delta$ where $\cf^{\HOD}(\kappa)=\cf(\kappa)$ and $\kappa^{+\HOD}=\kappa^{+}.$ In contrast, we also show that the $\HOD$ hypothesis is consistent with $\delta$ carrying a club of $\HOD$-regulars cardinals $\kappa$ such that $\kappa^{+\HOD}<\kappa^{+}$. Finally, we close the manuscript with a discussion about the $\HOD$ hypothesis and $\omega$-strong measurability.

The crux of quantum optics is using beam splitters to generate entanglement, including in pioneering experiments conducted by Hanbury-Brown and Twiss and Hong, Ou, and Mandel. This lies at the heart of what makes boson sampling hard to emulate by classical computers and is a vital component of quantum computation with light. Yet, despite overwhelming positive evidence, the conjecture that beam splitters with equal reflection and transmission probabilities generate the most entanglement for any state interfered with the vacuum has remained unproven for almost two decades [Asb\'oth et al., Phys. Rev. Lett. \textbf{94}, 173602 (2005)]. We prove this conjecture for ubiquitous entanglement monotones by uncovering monotonicity, convexity, and entropic properties of states undergoing photon loss. Because beam splitters are so fundamental, our results yield numerous corollaries for quantum optics, from inequalities for quasiprobability distributions to proofs of a recent conjecture for the evolution of a measure of quantumness through loss. One can now definitively state: the more balanced a beam splitter, the more entanglement it can generate with the vacuum., Comment: 5+5 pages, 1 figure, comments always welcome

Large language models (LLMs) represent a promising, but controversial, tool in aiding scientific peer review. This study evaluates the usefulness of LLMs in a conference setting as a tool for vetting paper submissions against submission standards. We conduct an experiment at the 2024 Neural Information Processing Systems (NeurIPS) conference, where 234 papers were voluntarily submitted to an "LLM-based Checklist Assistant." This assistant validates whether papers adhere to the author checklist used by NeurIPS, which includes questions to ensure compliance with research and manuscript preparation standards. Evaluation of the assistant by NeurIPS paper authors suggests that the LLM-based assistant was generally helpful in verifying checklist completion. In post-usage surveys, over 70% of authors found the assistant useful, and 70% indicate that they would revise their papers or checklist responses based on its feedback. While causal attribution to the assistant is not definitive, qualitative evidence suggests that the LLM contributed to improving some submissions. Survey responses and analysis of re-submissions indicate that authors made substantive revisions to their submissions in response to specific feedback from the LLM. The experiment also highlights common issues with LLMs: inaccuracy (20/52) and excessive strictness (14/52) were the most frequent issues flagged by authors. We also conduct experiments to understand potential gaming of the system, which reveal that the assistant could be manipulated to enhance scores through fabricated justifications, highlighting potential vulnerabilities of automated review tools.

An elementary embedding $j:M\rightarrow N$ between two inner models of ZFC is cardinal preserving if $M$ and $N$ correctly compute the class of cardinals. We look at the case $N=V$ and show that there is no nontrivial cardinal preserving elementary embedding from $M$ into $V$, answering a question of Caicedo.

In logistics, the ability to quickly compute and execute pick-and-place motions from bins is critical to increasing productivity. We present Bin-Optimized Motion Planning (BOMP), a motion planning framework that plans arm motions for a six-axis industrial robot with a long-nosed suction tool to remove boxes from deep bins. BOMP considers robot arm kinematics, actuation limits, the dimensions of a grasped box, and a varying height map of a bin environment to rapidly generate time-optimized, jerk-limited, and collision-free trajectories. The optimization is warm-started using a deep neural network trained offline in simulation with 25,000 scenes and corresponding trajectories. Experiments with 96 simulated and 15 physical environments suggest that BOMP generates collision-free trajectories that are up to 58 % faster than baseline sampling-based planners and up to 36 % faster than an industry-standard Up-Over-Down algorithm, which has an extremely low 15 % success rate in this context. BOMP also generates jerk-limited trajectories while baselines do not. Website: https://sites.google.com/berkeley.edu/bomp.

Text-to-image (T2I) models are remarkable at generating realistic images based on textual descriptions. However, textual prompts are inherently underspecified: they do not specify all possible attributes of the required image. This raises two key questions: Do T2I models generate diverse outputs on underspecified prompts? How can we automatically measure diversity? We propose GRADE: Granular Attribute Diversity Evaluation, an automatic method for quantifying sample diversity. GRADE leverages the world knowledge embedded in large language models and visual question-answering systems to identify relevant concept-specific axes of diversity (e.g., ``shape'' and ``color'' for the concept ``cookie''). It then estimates frequency distributions of concepts and their attributes and quantifies diversity using (normalized) entropy. GRADE achieves over 90% human agreement while exhibiting weak correlation to commonly used diversity metrics. We use GRADE to measure the overall diversity of 12 T2I models using 400 concept-attribute pairs, revealing that all models display limited variation. Further, we find that these models often exhibit default behaviors, a phenomenon where the model consistently generates concepts with the same attributes (e.g., 98% of the cookies are round). Finally, we demonstrate that a key reason for low diversity is due to underspecified captions in training data. Our work proposes a modern, semantically-driven approach to measure sample diversity and highlights the stunning homogeneity in outputs by T2I models., Comment: For project page and code see https://royira.github.io/GRADE

While numerous planetary and asteroid satellites show evidence for non-trivial rotation states, none are as emblematic as Hyperion, which has long been held as the most striking example of chaotic spin-orbit evolution in the Solar System. Nevertheless, an analytically tractable theory of the full 3D spin-orbit dynamics of Hyperion has not been developed. We derive the Hamiltonian for a spinning axisymmetric satellite in the gravitational potential of a planet without assuming planar or principal axis rotation and without averaging over the spin period. Using this model, we demonstrate the emergence of resonances between the nutation and orbital frequencies that act as the primary drivers of the spin dynamics. This analysis reveals that, contrary to long-held belief, Hyperion is not tumbling chaotically. Instead, it lies near or in a nutation-orbit resonance that is first-order in eccentricity, allowing it to rotate quasi-regularly. The most reliable observations are consistent with either nonchaotic motion or chaos that is orders of magnitude smaller than originally claimed. A separate phenomenon, the so-called barrel instability, is shown to be related to a different set of nutation-orbit resonances that generalize the planar spin-orbit resonances. Finally, we show that changes in spin states over long timescales are best understood by considering chaotic diffusion of quasi-conserved quantities., Comment: Accepted to A&A, 18 pages, 9 figures

We compute several types of dimension for the bounded derived categories of coherent sheaves of orbifold curves. This completes the calculation of these dimensions for derived categories of noncommutative curves in the sense of Reiten-van den Bergh. Along the way we construct stability conditions for orbifold curves. We also obtain a characterisation of orbifold curves with hereditary tilting bundle in terms of diagonal dimension., Comment: 23 pages, comments welcome!

Running optimization across many parallel seeds leveraging GPU compute have relaxed the need for a good initialization, but this can fail if the problem is highly non-convex as all seeds could get stuck in local minima. One such setting is collision-free motion optimization for robot manipulation, where optimization converges quickly on easy problems but struggle in obstacle dense environments (e.g., a cluttered cabinet or table). In these situations, graph-based planning algorithms are used to obtain seeds, resulting in significant slowdowns. We propose DiffusionSeeder, a diffusion based approach that generates trajectories to seed motion optimization for rapid robot motion planning. DiffusionSeeder takes the initial depth image observation of the scene and generates high quality, multi-modal trajectories that are then fine-tuned with a few iterations of motion optimization. We integrate DiffusionSeeder to generate the seed trajectories for cuRobo, a GPU-accelerated motion optimization method, which results in 12x speed up on average, and 36x speed up for more complicated problems, while achieving 10% higher success rate in partially observed simulation environments. Our results show the effectiveness of using diverse solutions from a learned diffusion model. Physical experiments on a Franka robot demonstrate the sim2real transfer of DiffusionSeeder to the real robot, with an average success rate of 86% and planning time of 26ms, improving on cuRobo by 51% higher success rate while also being 2.5x faster.

While coreference resolution is traditionally used as a component in individual document understanding, in this work we take a more global view and explore what can we learn about a domain from the set of all document-level coreference relations that are present in a large corpus. We derive coreference chains from a corpus of 30 million biomedical abstracts and construct a graph based on the string phrases within these chains, establishing connections between phrases if they co-occur within the same coreference chain. We then use the graph structure and the betweeness centrality measure to distinguish between edges denoting hierarchy, identity and noise, assign directionality to edges denoting hierarchy, and split nodes (strings) that correspond to multiple distinct concepts. The result is a rich, data-driven ontology over concepts in the biomedical domain, parts of which overlaps significantly with human-authored ontologies. We release the coreference chains and resulting ontology under a creative-commons license, along with the code.

We consider the problem of sampling from the ferromagnetic $q$-state Potts model on the random $d$-regular graph with parameter $\beta>0$. A key difficulty that arises in sampling from the model is the existence of a metastability window $(\beta_u,\beta_u')$ where the distribution has two competing modes, the so-called disordered and ordered phases, causing MCMC-based algorithms to be slow mixing from worst-case initialisations. To this end, Helmuth, Jenssen and Perkins designed a sampling algorithm that works for all $\beta$ when $q$ is large, using cluster expansion methods; more recently, their analysis technique has been adapted to show that random-cluster dynamics mixes fast when initialised more judiciously. However, a bottleneck behind cluster-expansion arguments is that they inherently only work for large $q$, whereas it is widely conjectured that sampling is possible for all $q,d\geq 3$. The only result so far that applies to general $q,d\geq 3$ is by Blanca and Gheissari who showed that the random-cluster dynamics mixes fast for $\beta<\beta_u$. For $\beta>\beta_u$, certain correlation phenomena emerge because of the metastability which have been hard to handle, especially for small $q$ and $d$. Our main contribution is to perform a delicate analysis of the Potts distribution and the random-cluster dynamics that goes beyond the threshold $\beta_u$. We use planting as the main tool in our proofs, and combine it with the analysis of random-cluster dynamics. We are thus able to show that the random-cluster dynamics initialised from all-out mixes fast for all integers $q,d\geq 3$ beyond the uniqueness threshold $\beta_u$; our analysis works all the way up to the threshold $\beta_c\in (\beta_u,\beta_u')$ where the dominant mode switches from disordered to ordered. We also obtain an algorithm in the ordered regime $\beta>\beta_c$ that refines significantly the range of $q,d$., Comment: Abstract shortened to meet arXiv requirements

In this paper, we showed that for suitable $(\beta,p, s,\ell)$ the $\beta$-order fractional derivative with respect to the last coordinate of the Fourier transform of an $L^p(\mathbb{R}^n)$ function is in $H^{-s}$ after restricting to a graph of a function with non-vanishing Gaussian curvature provided that the restriction of the Fourier transform of such function to the surface is in $H^{\ell}$. This is a generalization of the result in \cite{GoldStol}*{Theorem 1.12}., Comment: Fixed the description of Figure 3 and 4

Cloud robotics enables robots to offload computationally intensive tasks to cloud servers for performance, cost, and ease of management. However, the network and cloud computing infrastructure are not designed for reliable timing guarantees, due to fluctuating Quality-of-Service (QoS). In this work, we formulate an impossibility triangle theorem for: Latency reliability, Singleton server, and Commodity hardware. The LSC theorem suggests that providing replicated servers with uncorrelated failures can exponentially reduce the probability of missing a deadline. We present FogROS2-Probabilistic Latency Reliability (PLR) that uses multiple independent network interfaces to send requests to replicated cloud servers and uses the first response back. We design routing mechanisms to discover, connect, and route through non-default network interfaces on robots. FogROS2-PLR optimizes the selection of interfaces to servers to minimize the probability of missing a deadline. We conduct a cloud-connected driving experiment with two 5G service providers, demonstrating FogROS2-PLR effectively provides smooth service quality even if one of the service providers experiences low coverage and base station handover. We use 99 Percentile (P99) latency to evaluate anomalous long-tail latency behavior. In one experiment, FogROS2-PLR improves P99 latency by up to 3.7x compared to using one service provider. We deploy FogROS2-PLR on a physical Stretch 3 robot performing an indoor human-tracking task. Even in a fully covered Wi-Fi and 5G environment, FogROS2-PLR improves the responsiveness of the robot reducing mean latency by 36% and P99 latency by 33%., Comment: Submitted to 2025 IEEE International Conference on Robotics & Automation

Humans can learn to manipulate new objects by simply watching others; providing robots with the ability to learn from such demonstrations would enable a natural interface specifying new behaviors. This work develops Robot See Robot Do (RSRD), a method for imitating articulated object manipulation from a single monocular RGB human demonstration given a single static multi-view object scan. We first propose 4D Differentiable Part Models (4D-DPM), a method for recovering 3D part motion from a monocular video with differentiable rendering. This analysis-by-synthesis approach uses part-centric feature fields in an iterative optimization which enables the use of geometric regularizers to recover 3D motions from only a single video. Given this 4D reconstruction, the robot replicates object trajectories by planning bimanual arm motions that induce the demonstrated object part motion. By representing demonstrations as part-centric trajectories, RSRD focuses on replicating the demonstration's intended behavior while considering the robot's own morphological limits, rather than attempting to reproduce the hand's motion. We evaluate 4D-DPM's 3D tracking accuracy on ground truth annotated 3D part trajectories and RSRD's physical execution performance on 9 objects across 10 trials each on a bimanual YuMi robot. Each phase of RSRD achieves an average of 87% success rate, for a total end-to-end success rate of 60% across 90 trials. Notably, this is accomplished using only feature fields distilled from large pretrained vision models -- without any task-specific training, fine-tuning, dataset collection, or annotation. Project page: https://robot-see-robot-do.github.io, Comment: CoRL 2024, Project page: https://robot-see-robot-do.github.io

Building semantic 3D maps is valuable for searching for objects of interest in offices, warehouses, stores, and homes. We present a mapping system that incrementally builds a Language-Embedded Gaussian Splat (LEGS): a detailed 3D scene representation that encodes both appearance and semantics in a unified representation. LEGS is trained online as a robot traverses its environment to enable localization of open-vocabulary object queries. We evaluate LEGS on 4 room-scale scenes where we query for objects in the scene to assess how LEGS can capture semantic meaning. We compare LEGS to LERF and find that while both systems have comparable object query success rates, LEGS trains over 3.5x faster than LERF. Results suggest that a multi-camera setup and incremental bundle adjustment can boost visual reconstruction quality in constrained robot trajectories, and suggest LEGS can localize open-vocabulary and long-tail object queries with up to 66% accuracy.

Multiphoton interference is crucial to many photonic quantum technologies. In particular, interference forms the basis of optical quantum information processing platforms and can lead to significant computational advantages. It is therefore interesting to study the interference arising from various states of light in large interferometric networks. Here, we implement a quantum walk in a highly stable, low-loss, multiport interferometer with up to 24 ultrafast time bins. This time-bin interferometer comprises a sequence of birefringent crystals which produce pulses separated by 4.3\,ps, all along a single optical axis. Ultrafast Kerr gating in an optical fiber is employed to time-demultiplex the output from the quantum walk. We measure one-, two-, and three-photon interference arising from various input state combinations, including a heralded single-photon state, a thermal state, and an attenuated coherent state at one or more input ports. Our results demonstrate that ultrafast time bins are a promising platform to observe large-scale multiphoton interference.

Scaling up robot learning requires large and diverse datasets, and how to efficiently reuse collected data and transfer policies to new embodiments remains an open question. Emerging research such as the Open-X Embodiment (OXE) project has shown promise in leveraging skills by combining datasets including different robots. However, imbalances in the distribution of robot types and camera angles in many datasets make policies prone to overfit. To mitigate this issue, we propose RoVi-Aug, which leverages state-of-the-art image-to-image generative models to augment robot data by synthesizing demonstrations with different robots and camera views. Through extensive physical experiments, we show that, by training on robot- and viewpoint-augmented data, RoVi-Aug can zero-shot deploy on an unseen robot with significantly different camera angles. Compared to test-time adaptation algorithms such as Mirage, RoVi-Aug requires no extra processing at test time, does not assume known camera angles, and allows policy fine-tuning. Moreover, by co-training on both the original and augmented robot datasets, RoVi-Aug can learn multi-robot and multi-task policies, enabling more efficient transfer between robots and skills and improving success rates by up to 30%. Project website: https://rovi-aug.github.io., Comment: CoRL 2024 (Oral). Project website: https://rovi-aug.github.io

At the start of the COVID-19 pandemic, home visiting programs were faced with an unprecedented challenge: How do you deliver home visiting services without visiting homes? One home visiting model--Child First--quickly pivoted to telehealth, offering caregivers the option to receive services virtually. Child First has since resumed delivering services primarily in person, but some pandemic-driven implementation changes remain. To understand the impact of the pandemic on the Child First model, the study team conducted surveys and interviews with Child First staff members, and interviews with caregivers who received Child First services, to answer the following research questions: (1) To what extent did the implementation of Child First services change since the start of the pandemic?; (2) How did Child First staff members report implementing core components of the model since the pandemic began?; and (3) How did Child First caregivers report on Child First services that they received since the pandemic began? Overall, the study team found that the implementation of the Child First model following the first three years of the pandemic remains largely consistent with pre-pandemic implementation, despite the unique challenges to home visiting posed by the pandemic.