Search

Cancer survivors experience a wide range of impairments arising from cancer or its treatment, such as chemo brain, visual impairments, and physical impairments. These impairments degrade their quality of life and potentially make software use more challenging for them. However, there has been limited research on designing accessible software for cancer survivors. To bridge this research gap, we conducted a formative study including a survey (n=46), semi-structured interviews (n=20), and a diary study (n=10) with cancer survivors. Our results revealed a wide range of impairments experienced by cancer survivors, including chemo brain, neuropathy, and visual impairments. Cancer survivors heavily relied on software for socialization, health purposes, and cancer advocacy, but their impairments made software use more challenging for them. Based on the results, we offer a set of accessibility guidelines that software designers can utilize when creating applications for cancer survivors. Further, we suggest design features for inclusion, such as health resources, socialization tools, and games, tailored to the needs of cancer survivors. This research aims to spotlight cancer survivors' software accessibility challenges and software needs and invite more research in this important yet under-investigated domain.

In recent years, there has been a trend in the field of Reinforcement Learning (RL) towards large action models trained offline on large-scale datasets via sequence modeling. Existing models are primarily based on the Transformer architecture, which result in powerful agents. However, due to slow inference times, Transformer-based approaches are impractical for real-time applications, such as robotics. Recently, modern recurrent architectures, such as xLSTM and Mamba, have been proposed that exhibit parallelization benefits during training similar to the Transformer architecture while offering fast inference. In this work, we study the aptitude of these modern recurrent architectures for large action models. Consequently, we propose a Large Recurrent Action Model (LRAM) with an xLSTM at its core that comes with linear-time inference complexity and natural sequence length extrapolation abilities. Experiments on 432 tasks from 6 domains show that LRAM compares favorably to Transformers in terms of performance and speed.

The interplay of spin and charge degrees of freedom is believed to underlie various unresolved phenomena in strongly correlated systems. Quantum simulators based on neutral atoms provide an excellent testbed for investigating such phenomena and resolving their microscopic origins. Up to now, the majority of experimental and theoretical studies has focused on systems with fermionic exchange statistics. Here we expand the existing cold atom toolbox through the use of negative temperature states, enabling us to realize an antiferromagnetic, bosonic $t-J$ model in two spatial dimensions, subject to a strong staggered magnetic field in a quantum gas microscope. Through comparison of the spreading dynamics of a single hole in a N\'eel versus a spin-polarized initial state, we establish the relevance of memory effects resulting from the buildup of strong spin-charge correlations in the dynamics of charge carriers in antiferromagnets. We further numerically predict rich dynamics of pairs of doped holes, which we demonstrate to be bound by a similar memory effect, while their center-of-mass can expand freely. Our work paves the way for the systematic exploration of the effect of antiferromagnetic spin ordering on the properties of individual charge carriers as well as finite doping phases: Our study demonstrates that the staggered field can be used to single out the effect of antiferromagnetism and holds the prospect to prepare low-temperature states in the near future., Comment: 9+3 pages, 4+5 figures

We describe the cryogenic system being developed for Taurus: a super-pressure balloon-borne microwave polarimeter scheduled to fly in 2027. The Taurus cryogenic system consists of a 660L liquid helium cryostat which achieves a base temperature of <100mK with the help of a capillary-fed superfluid tank and a closed cycle dilution refrigerator. The main tank is supported with fiberglass flexures and is encased in two layers of vapor-cooled shields which allow Taurus to make full use of the extended flight time offered by the super-pressure balloon platform. The Taurus cryostat is projected to hold for over 50 days while weighing under 1000lbs. We present the design, testing, and thermal analysis of the Taurus cryogenic systems.

Large language models (LLMs) acquire beliefs about gender from training data and can therefore generate text with stereotypical gender attitudes. Prior studies have demonstrated model generations favor one gender or exhibit stereotypes about gender, but have not investigated the complex dynamics that can influence model reasoning and decision-making involving gender. We study gender equity within LLMs through a decision-making lens with a new dataset, DeMET Prompts, containing scenarios related to intimate, romantic relationships. We explore nine relationship configurations through name pairs across three name lists (men, women, neutral). We investigate equity in the context of gender roles through numerous lenses: typical and gender-neutral names, with and without model safety enhancements, same and mixed-gender relationships, and egalitarian versus traditional scenarios across various topics. While all models exhibit the same biases (women favored, then those with gender-neutral names, and lastly men), safety guardrails reduce bias. In addition, models tend to circumvent traditional male dominance stereotypes and side with 'traditionally female' individuals more often, suggesting relationships are viewed as a female domain by the models., Comment: EMNLP Findings 2024

The equivalence between von Neumann's Minimax Theorem for zero-sum games and the LP Duality Theorem connects cornerstone problems of the two fields of game theory and optimization, respectively, and has been the subject of intense scrutiny for seven decades. Yet, as observed in this paper, the proof of the difficult direction of this equivalence is unsatisfactory: It does not assign distinct roles to the two players of the game, as is natural from the definition of a zero-sum game. In retrospect, a partial resolution to this predicament was provided in another brilliant paper of von Neumann (1953), which reduced the assignment problem to zero-sum games. However, the underlying LP is highly specialized; all entries of its objective function vector are strictly positive and all entries of the constraint matrix and right hand side vector are equal to one. We generalize von Neumann's result along two directions, each allowing negative entries in certain parts of the LP. Our reductions make explicit the roles of the two players of the reduced game, namely their maximin strategies are to play optimal solutions to the primal and dual LPs. Furthermore, unlike previous reductions, the value of the reduced game reveals the value of the given LP. Our generalizations encompass several basic economic scenarios. We end by discussing evidence that von Neumann possessed an understanding of the notion of polynomial-time solvability.

We investigate the magnetic quantum phase-transitions in bulk correlated metals at the level of dynamical mean-field theory. To this end, we focus on the Hubbard model on a simple cubic lattice as a function of temperature and electronic density, determining the different regimes of its magnetic transition - classical, quantum critical, and quantum disordered - as well as the corresponding critical exponents. Our numerical results, together with supporting mean-field derivations, demonstrate how the presence of Kohn-anomalies on the underlying Fermi surface does not only drive the quantum critical behavior above the quantum critical point, but shapes the whole phase diagram around it. Finally, after outlining the impact of different Fermi surface geometries on quantum criticality, we discuss to what extent spatial correlations beyond dynamical mean-field might modify our findings., Comment: 7 pages (including references), 3 figures

Recent advances in the understanding of neural networks suggest that superposition, the ability of a single neuron to represent multiple features simultaneously, is a key mechanism underlying the computational efficiency of large-scale networks. This paper explores the theoretical foundations of computing in superposition, focusing on explicit, provably correct algorithms and their efficiency. We present the first lower bounds showing that for a broad class of problems, including permutations and pairwise logical operations, a neural network computing in superposition requires at least $\Omega(m' \log m')$ parameters and $\Omega(\sqrt{m' \log m'})$ neurons, where $m'$ is the number of output features being computed. This implies that any ``lottery ticket'' sparse sub-network must have at least $\Omega(m' \log m')$ parameters no matter what the initial dense network size. Conversely, we show a nearly tight upper bound: logical operations like pairwise AND can be computed using $O(\sqrt{m'} \log m')$ neurons and $O(m' \log^2 m')$ parameters. There is thus an exponential gap between computing in superposition, the subject of this work, and representing features in superposition, which can require as little as $O(\log m'$) neurons based on the Johnson-Lindenstrauss Lemma. Our hope is that our results open a path for using complexity theoretic techniques in neural network interpretability research., Comment: 43 pages, 8 figures

Anonymity is an important principle online. However, malicious actors have long used misleading identities to conduct fraud, spread disinformation, and carry out other deceptive schemes. With the advent of increasingly capable AI, bad actors can amplify the potential scale and effectiveness of their operations, intensifying the challenge of balancing anonymity and trustworthiness online. In this paper, we analyze the value of a new tool to address this challenge: "personhood credentials" (PHCs), digital credentials that empower users to demonstrate that they are real people -- not AIs -- to online services, without disclosing any personal information. Such credentials can be issued by a range of trusted institutions -- governments or otherwise. A PHC system, according to our definition, could be local or global, and does not need to be biometrics-based. Two trends in AI contribute to the urgency of the challenge: AI's increasing indistinguishability from people online (i.e., lifelike content and avatars, agentic activity), and AI's increasing scalability (i.e., cost-effectiveness, accessibility). Drawing on a long history of research into anonymous credentials and "proof-of-personhood" systems, personhood credentials give people a way to signal their trustworthiness on online platforms, and offer service providers new tools for reducing misuse by bad actors. In contrast, existing countermeasures to automated deception -- such as CAPTCHAs -- are inadequate against sophisticated AI, while stringent identity verification solutions are insufficiently private for many use-cases. After surveying the benefits of personhood credentials, we also examine deployment risks and design challenges. We conclude with actionable next steps for policymakers, technologists, and standards bodies to consider in consultation with the public., Comment: 63 pages, 7 figures, 5 tables; minor additions to acknowledgments and wording changes for clarity; corrected typo

Let $F$ be a nonarchimedean local field of residual characteristic $p$. Let $G$ denote a connected reductive group over $F$ that splits over a tamely ramified extension of $F$. Let $(K ,\rho)$ be a type as constructed by Kim and Yu. We show that there exists a twisted Levi subgroup $G^0 \subset G$ and a type $(K^0, \rho^0)$ for $G^0$ such that the corresponding Hecke algebras $\mathcal{H}(G(F), (K, \rho))$ and $\mathcal{H}(G^0(F), (K^0, \rho^0))$ are isomorphic. If $p$ does not divide the order of the absolute Weyl group of $G$, then every Bernstein block is equivalent to modules over such a Hecke algebra. Hence, under this assumption on $p$, our result implies that every Bernstein block is equivalent to a depth-zero Bernstein block. This allows one to reduce many problems about (the category of) smooth, complex representations of $p$-adic groups to analogous problems about (the category of) depth-zero representations. Our isomorphism of Hecke algebras is very explicit and also includes an explicit description of the Hecke algebras as semi-direct products of an affine Hecke with a twisted group algebra. Moreover, we work with arbitrary algebraically closed fields of characteristic different from $p$ as our coefficient field. This paper relies on a prior axiomatic result about the structure of Hecke algebras by the same authors and a key ingredient consists of extending the quadratic character of Fintzen--Kaletha--Spice to the support of the Hecke algebra, which might be of independent interest., Comment: 62 pages; this paper relies on a prior paper by the same by the same authors mentioned in the abstract and submitted to the arxiv at the same time, we recommend saving both papers in the same folder (saving the present paper as Adler--Fintzen--Mishra--Ohara_Reduction_to_depth_zero_for_tame_p-adic_groups_via_Hecke_algebra_isomorphisms.pdf) to take advantage of the hyperlinks between them

Let $G$ denote a connected reductive group over a nonarchimedean local field $F$ of residue characteristic $p$, and let $\mathcal{C}$ denote an algebraically closed field of characteristic $\ell \neq p$. If $\rho$ is an irreducible, smooth $\mathcal{C}$-representation of a compact, open subgroup $K$ of $G(F)$, then the pair $(K,\rho)$ gives rise to a Hecke algebra $\mathcal{H}(G(F),(K, \rho))$. For a large class of pairs $(K,\rho)$, we show that $\mathcal{H}(G(F),(K, \rho))$ is a semi-direct product of an affine Hecke algebra with explicit parameters with a twisted group algebra, and that it is isomorphic to $\mathcal{H}(G^0(F),(K^0, \rho^0))$ for some reductive subgroup $G^0 \subset G$ with compact, open subgroup $K^0$ and depth-zero representation $\rho^0$ of $K^0$. The class of pairs that we consider includes all depth-zero types. In describing their Hecke algebras, we thus recover a result of Morris as a special case. In a second paper, we will show that our class also contains all the types constructed by Kim and Yu, and hence we obtain as a corollary that arbitrary Bernstein blocks are equivalent to depth-zero Bernstein blocks under minor tameness assumptions. The pairs to which our results apply are described in an axiomatic way so that the results can be applied to other constructions of types by only verifying that the relevant axioms are satisfied. The Hecke algebra isomorphisms are given in an explicit manner and are support preserving., Comment: 87 pages; this paper contains a lot of hyperlinks to the sequel paper by the same authors mentioned in the abstract and submitted to the arxiv at the same time, we recommend saving both papers in the same folder (saving the present paper as Adler--Fintzen--Mishra--Ohara_Structure_of_Hecke_algebras_arising_from_types.pdf) to take advantage of these hyperlinks

In this work, we present a new stabilization method aimed at removing spurious oscillations in the pressure approximation of Biot's model for poroelasticity with low permeabilities and/or small time steps. We consider different finite-element discretizations and illustrate how not only does such a stabilized scheme provide numerical solutions that are free of non-physical oscillations, but it also allows one to iterate the fluid and mechanics problems in a fashion similar to the well-known fixed-stress split method. The resulting solution method is convergent without the necessity for additional terms to stabilize the iteration. Finally, we present numerical results illustrating the robust behavior of both the stabilization and iterative solver with respect to the physical and discretization parameters of the model.

Machine learning techniques have recently been of great interest for solving differential equations. Training these models is classically a data-fitting task, but knowledge of the expression of the differential equation can be used to supplement the training objective, leading to the development of physics-informed scientific machine learning. In this article, we focus on one class of models called nonlinear vector autoregression (NVAR) to solve ordinary differential equations (ODEs). Motivated by connections to numerical integration and physics-informed neural networks, we explicitly derive the physics-informed NVAR (piNVAR) which enforces the right-hand side of the underlying differential equation regardless of NVAR construction. Because NVAR and piNVAR completely share their learned parameters, we propose an augmented procedure to jointly train the two models. Then, using both data-driven and ODE-driven metrics, we evaluate the ability of the piNVAR model to predict solutions to various ODE systems, such as the undamped spring, a Lotka-Volterra predator-prey nonlinear model, and the chaotic Lorenz system.

When developing robust preconditioners for multiphysics problems, fractional functions of the Laplace operator often arise and need to be inverted. Rational approximation in the uniform norm can be used to convert inverting those fractional operators into inverting a series of shifted Laplace operators. Care must be taken in the approximation so that the shifted Laplace operators remain symmetric positive definite, making them better conditioned. In this work, we study two greedy algorithms for finding rational approximations to such fractional operators. The first algorithm improves the orthogonal greedy algorithm discussed in [Li et al., SISC, 2024] by adding one minimization step in the uniform norm to the procedure. The second approach employs the weak Chebyshev greedy algorithm in the uniform norm. Both methods yield non-increasing error. Numerical results confirm the effectiveness of our proposed algorithms, which are also flexible and applicable to other approximation problems. Moreover, with effective rational approximations to the fractional operator, the resulting algorithms show good performance in preconditioning a Darcy-Stokes coupled problem.

Large Language Models (LLMs) are able to provide assistance on a wide range of information-seeking tasks. However, model outputs may be misleading, whether unintentionally or in cases of intentional deception. We investigate the ability of LLMs to be deceptive in the context of providing assistance on a reading comprehension task, using LLMs as proxies for human users. We compare outcomes of (1) when the model is prompted to provide truthful assistance, (2) when it is prompted to be subtly misleading, and (3) when it is prompted to argue for an incorrect answer. Our experiments show that GPT-4 can effectively mislead both GPT-3.5-Turbo and GPT-4, with deceptive assistants resulting in up to a 23% drop in accuracy on the task compared to when a truthful assistant is used. We also find that providing the user model with additional context from the passage partially mitigates the influence of the deceptive model. This work highlights the ability of LLMs to produce misleading information and the effects this may have in real-world situations., Comment: Next Generation of AI Safety Workshop, 41st International Conference on Machine Learning (ICML 2024)

We study the 3D-consistency property for negative symmetries of KdV type equations. Its connection with the 3D-consistency of discrete equations is explained., Comment: 19 pages

In recent years, linear complementary pairs (LCP) of codes and linear complementary dual (LCD) codes have gained significant attention due to their applications in coding theory and cryptography. In this work, we construct explicit LCPs of codes and LCD codes from function fields of genus $g \geq 1$. To accomplish this, we present pairs of suitable divisors giving rise to non-special divisors of degree $g-1$ in the function field. The results are applied in constructing LCPs of algebraic geometry codes and LCD algebraic geometry (AG) codes in Kummer extensions, hyperelliptic function fields, and elliptic curves.

Taurus is a balloon-borne cosmic microwave background (CMB) experiment optimized to map the E-mode polarization and Galactic foregrounds at the largest angular scales ($\ell$ $\lt$ 30) and improve measurements of the optical depth to reionization ($\tau$). This will pave the way for improved measurements of the sum of neutrino masses in combination with high-resolution CMB data while also testing the $\Lambda CDM$ model on large angular scales and providing high-frequency maps of polarized dust foregrounds to the CMB community. These measurements take advantage of the low-loading environment found in the stratosphere and are enabled by NASA's super-pressure balloon platform, which provides access to 70% of the sky with a launch from Wanaka, New Zealand. Here we describe a general overview of Taurus, with an emphasis on the instrument design. Taurus will employ more than 10,000 100 mK transition edge sensor bolometers distributed across two low-frequency (150, 220 GHz) and one high-frequency (280, 350 GHz) dichroic receivers. The liquid helium cryostat housing the detectors and optics is supported by a lightweight gondola. The payload is designed to meet the challenges in mass, power, and thermal control posed by the super-pressure platform. The instrument and scan strategy are optimized for rigorous control of instrumental systematics, enabling high-fidelity linear polarization measurements on the largest angular scales.

We simulate a variety of optical systematics for Taurus, a balloon-borne cosmic microwave background (CMB) polarisation experiment, to assess their impact on large-scale E-mode polarisation measurements and constraints of the optical depth to reionisation {\tau}. We model a one-month flight of Taurus from Wanaka, New Zealand aboard a super-pressure balloon (SPB). We simulate night-time scans of both the CMB and dust foregrounds in the 150GHz band, one of Taurus's four observing bands. We consider a variety of possible systematics that may affect Taurus's observations, including non-gaussian beams, pointing reconstruction error, and half-wave plate (HWP) non-idealities. For each of these, we evaluate the residual power in the difference between maps simulated with and without the systematic, and compare this to the expected signal level corresponding to Taurus's science goals. Our results indicate that most of the HWP-related systematics can be mitigated to be smaller than sample variance by calibrating with Planck's TT spectrum and using an achromatic HWP model, with a preference for five layers of sapphire to ensure good systematic control. However, additional beam characterization will be required to mitigate far-sidelobe pickup from dust on larger scales., Comment: 21 pages, 6 tables, 7 figures, this version as published in JCAP after minor revisions to v1

We release the Nemotron-4 340B model family, including Nemotron-4-340B-Base, Nemotron-4-340B-Instruct, and Nemotron-4-340B-Reward. Our models are open access under the NVIDIA Open Model License Agreement, a permissive model license that allows distribution, modification, and use of the models and its outputs. These models perform competitively to open access models on a wide range of evaluation benchmarks, and were sized to fit on a single DGX H100 with 8 GPUs when deployed in FP8 precision. We believe that the community can benefit from these models in various research studies and commercial applications, especially for generating synthetic data to train smaller language models. Notably, over 98% of data used in our model alignment process is synthetically generated, showcasing the effectiveness of these models in generating synthetic data. To further support open research and facilitate model development, we are also open-sourcing the synthetic data generation pipeline used in our model alignment process.

Constructing the Burrows-Wheeler transform (BWT) for long strings poses significant challenges regarding construction time and memory usage. We use a prefix of the suffix array to partition a long string into shorter substrings, thereby enabling the use of multi-string BWT construction algorithms to process these partitions fast. We provide an implementation partDNA for DNA sequences. Through comparison with state-of-the-art BWT construction algorithms, we show that partDNA with ropebwt2 offers a novel trade-off for construction time and memory usage for BWT construction on real genome datasets. Beyond this, the proposed partitioning strategy is applicable to strings of any alphabet.

Magnetic resonance imaging (MRI) is a crucial tool to identify brain abnormalities in a wide range of neurological disorders. In focal epilepsy MRI is used to identify structural cerebral abnormalities. For covert lesions, machine learning and artificial intelligence algorithms may improve lesion detection if abnormalities are not evident on visual inspection. The success of this approach depends on the volume and quality of training data. Herein, we release an open-source dataset of preprocessed MRI scans from 442 individuals with drug-refractory focal epilepsy who had neurosurgical resections, and detailed demographic information. The MRI scan data includes the preoperative 3D T1 and where available 3D FLAIR, as well as a manually inspected complete surface reconstruction and volumetric parcellations. Demographic information includes age, sex, age of onset of epilepsy, location of surgery, histopathology of resected specimen, occurrence and frequency of focal seizures with and without impairment of awareness, focal to bilateral tonic-clonic seizures, number of anti-seizure medications (ASMs) at time of surgery, and a total of 1764 patient years of post-surgical follow up. Crucially, we also include resection masks delineated from post-surgical imaging. To demonstrate the veracity of our data, we successfully replicated previous studies showing long-term outcomes of seizure freedom in the range of around 50%. Our imaging data replicates findings of group level atrophy in patients compared to controls. Resection locations in the cohort were predominantly in the temporal and frontal lobes. We envisage our dataset, shared openly with the community, will catalyse the development and application of computational methods in clinical neurology.

Accessibility education has been rarely incorporated into the high school curricula. This is a missed opportunity to equip next-generation software designers and decision-makers with knowledge, awareness, and empathy regarding accessibility and disabilities. We taught accessibility to students (N=93) in a midwestern high school through empathy-driven games and interviewed three Computer Science high school teachers and one librarian who taught programming. Accessibility education is currently insufficient in high school, facing challenges such as teachers' knowledge and conflicted curriculum goals. The students exhibited increased knowledge and awareness of accessibility and empathy for people with disabilities after playing the games. With this education outreach, we aim to provide insights into teaching next-generation software designers about accessibility by leveraging games., Comment: 87th Annual Meeting of the Association for Information Science and Technology (ASIS&T)

We present a charge self-consistent density functional theory combined with the ghost-rotationally-invariant slave-boson (DFT+gRISB) formalism for studying correlated materials. This method is applied to SrVO$_3$ and NiO, representing prototypical correlated metals and charge-transfer insulators. For SrVO$_3$, we demonstrate that DFT+gRISB yields an accurate equilibrium volume and effective mass close to experimentally observed values. Regarding NiO, DFT+gRISB enables the simultaneous description of charge transfer and Mott-Hubbard bands, significantly enhancing the accuracy of the original DFT+RISB approach. Furthermore, the calculated equilibrium volume and spectral function reasonably agree with experimental observations.

LiteBIRD, the next-generation cosmic microwave background (CMB) experiment, aims for a launch in Japan's fiscal year 2032, marking a major advancement in the exploration of primordial cosmology and fundamental physics. Orbiting the Sun-Earth Lagrangian point L2, this JAXA-led strategic L-class mission will conduct a comprehensive mapping of the CMB polarization across the entire sky. During its 3-year mission, LiteBIRD will employ three telescopes within 15 unique frequency bands (ranging from 34 through 448 GHz), targeting a sensitivity of 2.2\,$\mu$K-arcmin and a resolution of 0.5$^\circ$ at 100\,GHz. Its primary goal is to measure the tensor-to-scalar ratio $r$ with an uncertainty $\delta r = 0.001$, including systematic errors and margin. If $r \geq 0.01$, LiteBIRD expects to achieve a $>5\sigma$ detection in the $\ell=$2-10 and $\ell=$11-200 ranges separately, providing crucial insight into the early Universe. We describe LiteBIRD's scientific objectives, the application of systems engineering to mission requirements, the anticipated scientific impact, and the operations and scanning strategies vital to minimizing systematic effects. We will also highlight LiteBIRD's synergies with concurrent CMB projects., Comment: 23 pages, 9 figures, 1 table, SPIE Astronomical Telescopes + Instrumentation 2024

The competition of different length scales in quantum many-body systems leads to various novel phenomena, including the emergence of correlated dynamics or non-local order. To access and investigate such effects in an itinerant lattice-based quantum simulator, it has been proposed to introduce tunable extended-range interactions using off-resonant optical coupling to Rydberg states. However, experimental realizations of such "Rydberg dressing" have so far mostly concentrated on spin systems without motion. Here, we overcome a number of experimental challenges limiting previous work and realize an effective one-dimensional extended Bose-Hubbard model (eBHM). Harnessing our quantum gas microscope, we probe the correlated out-of-equilibrium dynamics of extended-range repulsively-bound pairs at low filling, and kinetically-constrained "hard rods" at half filling. Near equilibrium, we observe density ordering when adiabatically turning on the extended-range interactions. Our results demonstrate the versatility of Rydberg dressing in engineering itinerant optical lattice-based quantum simulators and pave the way to realizing novel light-controlled extended-range interacting quantum many-body systems., Comment: 17 pages, 12 figures

The relaxation behaviour of isolated quantum systems taken out of equilibrium is among the most intriguing questions in many-body physics. Quantum systems out of equilibrium typically relax to thermal equilibrium states by scrambling local information and building up entanglement entropy. However, kinetic constraints in the Hamiltonian can lead to a breakdown of this fundamental paradigm due to a fragmentation of the underlying Hilbert space into dynamically decoupled subsectors in which thermalisation can be strongly suppressed. Here, we experimentally observe Hilbert space fragmentation (HSF) in a two-dimensional tilted Bose-Hubbard model. Using quantum gas microscopy, we engineer a wide variety of initial states and find a rich set of manifestations of HSF involving bulk states, interfaces and defects, i.e., d = 2, 1 and 0 dimensional objects. Specifically, uniform initial states with equal particle number and energy differ strikingly in their relaxation dynamics. Inserting controlled defects on top of a global, non-thermalising chequerboard state, we observe highly anisotropic, sub-dimensional dynamics, an immediate signature of their fractonic nature. An interface between localized and thermalising states in turn displays dynamics depending on its orientation. Our results mark the first observation of HSF beyond one dimension, as well as the concomitant direct observation of fractons, and pave the way for in-depth studies of microscopic transport phenomena in constrained systems, Comment: 14 pages, 12 figures

Tourism affects not only the tourism industry but also society and stakeholders such as the environment, local businesses, and residents. Tourism Recommender Systems (TRS) can be pivotal in promoting sustainable tourism by guiding travelers toward destinations with minimal negative impact. Our paper introduces a composite sustainability indicator for a city trip TRS based on the users' starting point and month of travel. This indicator integrates CO2e emissions for different transportation modes and analyses destination popularity and seasonal demand. We quantify city popularity based on user reviews, points of interest, and search trends from Tripadvisor and Google Trends data. To calculate a seasonal demand index, we leverage data from TourMIS and Airbnb. We conducted a user study to explore the fundamental trade-offs in travel decision-making and determine the weights for our proposed indicator. Finally, we demonstrate the integration of this indicator into a TRS, illustrating its ability to deliver sustainable city trip recommendations. This work lays the foundation for future research by integrating sustainability measures and contributing to responsible recommendations by TRS.

Meta-reinforcement learning (meta-RL) is a promising framework for tackling challenging domains requiring efficient exploration. Existing meta-RL algorithms are characterized by low sample efficiency, and mostly focus on low-dimensional task distributions. In parallel, model-based RL methods have been successful in solving partially observable MDPs, of which meta-RL is a special case. In this work, we leverage this success and propose a new model-based approach to meta-RL, based on elements from existing state-of-the-art model-based and meta-RL methods. We demonstrate the effectiveness of our approach on common meta-RL benchmark domains, attaining greater return with better sample efficiency (up to $15\times$) while requiring very little hyperparameter tuning. In addition, we validate our approach on a slate of more challenging, higher-dimensional domains, taking a step towards real-world generalizing agents.

In this report, we introduce the Gemini 1.5 family of models, representing the next generation of highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio. The family includes two new models: (1) an updated Gemini 1.5 Pro, which exceeds the February version on the great majority of capabilities and benchmarks; (2) Gemini 1.5 Flash, a more lightweight variant designed for efficiency with minimal regression in quality. Gemini 1.5 models achieve near-perfect recall on long-context retrieval tasks across modalities, improve the state-of-the-art in long-document QA, long-video QA and long-context ASR, and match or surpass Gemini 1.0 Ultra's state-of-the-art performance across a broad set of benchmarks. Studying the limits of Gemini 1.5's long-context ability, we find continued improvement in next-token prediction and near-perfect retrieval (>99%) up to at least 10M tokens, a generational leap over existing models such as Claude 3.0 (200k) and GPT-4 Turbo (128k). Finally, we highlight real-world use cases, such as Gemini 1.5 collaborating with professionals on completing their tasks achieving 26 to 75% time savings across 10 different job categories, as well as surprising new capabilities of large language models at the frontier; when given a grammar manual for Kalamang, a language with fewer than 200 speakers worldwide, the model learns to translate English to Kalamang at a similar level to a person who learned from the same content.

Chat-based large language models have the opportunity to empower individuals lacking high-quality healthcare access to receive personalized information across a variety of topics. However, users may ask underspecified questions that require additional context for a model to correctly answer. We study how large language model biases are exhibited through these contextual questions in the healthcare domain. To accomplish this, we curate a dataset of sexual and reproductive healthcare questions that are dependent on age, sex, and location attributes. We compare models' outputs with and without demographic context to determine group alignment among our contextual questions. Our experiments reveal biases in each of these attributes, where young adult female users are favored.