Search

Across various applications, humans increasingly use black-box artificial intelligence (AI) systems without insight into these systems' reasoning. To counter this opacity, explainable AI (XAI) methods promise enhanced transparency and interpretability. While recent studies have explored how XAI affects human-AI collaboration, few have examined the potential pitfalls caused by incorrect explanations. The implications for humans can be far-reaching but have not been explored extensively. To investigate this, we ran a study (n=160) on AI-assisted decision-making in which humans were supported by XAI. Our findings reveal a misinformation effect when incorrect explanations accompany correct AI advice with implications post-collaboration. This effect causes humans to infer flawed reasoning strategies, hindering task execution and demonstrating impaired procedural knowledge. Additionally, incorrect explanations compromise human-AI team-performance during collaboration. With our work, we contribute to HCI by providing empirical evidence for the negative consequences of incorrect explanations on humans post-collaboration and outlining guidelines for designers of AI.

In this work, we present MedImageInsight, an open-source medical imaging embedding model. MedImageInsight is trained on medical images with associated text and labels across a diverse collection of domains, including X-Ray, CT, MRI, dermoscopy, OCT, fundus photography, ultrasound, histopathology, and mammography. Rigorous evaluations demonstrate MedImageInsight's ability to achieve state-of-the-art (SOTA) or human expert level performance across classification, image-image search, and fine-tuning tasks. Specifically, on public datasets, MedImageInsight achieves SOTA in CT 3D medical image retrieval, as well as SOTA in disease classification and search for chest X-ray, dermatology, and OCT imaging. Furthermore, MedImageInsight achieves human expert performance in bone age estimation (on both public and partner data), as well as AUC above 0.9 in most other domains. When paired with a text decoder, MedImageInsight achieves near SOTA level single image report findings generation with less than 10\% the parameters of other models. Compared to fine-tuning GPT-4o with only MIMIC-CXR data for the same task, MedImageInsight outperforms in clinical metrics, but underperforms on lexical metrics where GPT-4o sets a new SOTA. Importantly for regulatory purposes, MedImageInsight can generate ROC curves, adjust sensitivity and specificity based on clinical need, and provide evidence-based decision support through image-image search (which can also enable retrieval augmented generation). In an independent clinical evaluation of image-image search in chest X-ray, MedImageInsight outperformed every other publicly available foundation model evaluated by large margins (over 6 points AUC), and significantly outperformed other models in terms of AI fairness (across age and gender). We hope releasing MedImageInsight will help enhance collective progress in medical imaging AI research and development.

We show that Koszul duality between differential graded categories and pointed curved coalgebras interchanges smooth and proper Calabi-Yau structures. This result is a generalization and conceptual explanation of the following two applications. For a finite-dimensional Lie algebra a smooth Calabi-Yau structure on the universal enveloping algebra is equivalent to a proper Calabi-Yau structure on the Chevalley-Eilenberg chain coalgebra, which exists if and only if Poincare duality is satisfied. For a topological space X having the homotopy type of a finite complex we show an oriented Poincare duality structure (with local coefficients) on X is equivalent to a proper Calabi-Yau structure on the dg coalgebra of chains on X and to a smooth Calabi-Yau structure on the dg algebra of chains on the based loop space of X., Comment: V2: Minor clarifications and corrections. 32 pages

Whether a large language model policy is an explicit constitution or an implicit reward model, it is challenging to assess coverage over the unbounded set of real-world situations that a policy must contend with. We introduce an AI policy design process inspired by mapmaking, which has developed tactics for visualizing and iterating on maps even when full coverage is not possible. With Policy Projector, policy designers can survey the landscape of model input-output pairs, define custom regions (e.g., "violence"), and navigate these regions with rules that can be applied to LLM outputs (e.g., if output contains "violence" and "graphic details," then rewrite without "graphic details"). Policy Projector supports interactive policy authoring using LLM classification and steering and a map visualization reflecting the policy designer's work. In an evaluation with 12 AI safety experts, our system helps policy designers to address problematic model behaviors extending beyond an existing, comprehensive harm taxonomy.

Community and organizational policies are typically designed in a top-down, centralized fashion, with limited input from impacted stakeholders. This can result in policies that are misaligned with community needs or perceived as illegitimate. How can we support more collaborative, participatory approaches to policy design? In this paper, we present PolicyCraft, a system that structures collaborative policy design through case-grounded deliberation. Building on past research that highlights the value of concrete cases in establishing common ground, PolicyCraft supports users in collaboratively proposing, critiquing, and revising policies through discussion and voting on cases. A field study across two university courses showed that students using PolicyCraft reached greater consensus and developed better-supported course policies, compared with those using a baseline system that did not scaffold their use of concrete cases. Reflecting on our findings, we discuss opportunities for future HCI systems to help groups more effectively bridge between abstract policies and concrete cases.

The increasing use of generative Artificial Intelligence (AI) in modern software engineering, particularly Large Language Models (LLMs) for code generation, has transformed professional software development by boosting productivity and automating development processes. This adoption, however, has highlighted a significant issue: the introduction of security vulnerabilities into the code. These vulnerabilities result, e.g., from flaws in the training data that propagate into the generated code, creating challenges in disclosing them. Traditional vulnerability handling processes often involve extensive manual review. Applying such traditional processes to AI-generated code is challenging. AI-generated code may include several vulnerabilities, possibly in slightly different forms as developers might not build on already implemented code but prompt similar tasks. In this work, we explore the current state of LLM-based approaches for vulnerability handling, focusing on approaches for vulnerability detection, localization, and repair. We provide an overview of recent progress in this area and highlight open challenges that must be addressed in order to establish a reliable and scalable vulnerability handling process of AI-generated code., Comment: Accepted for publication @ IEEE AIxSET 2024; 4 pages, 2 Figures

Let $X$ be a (-1)-shifted symplectic derived Deligne--Mumford stack. In this paper we introduce the Darboux stack of $X$, parametrizing local presentations of $X$ as a derived critical locus of a function $f$ on a smooth formal scheme $U$. Local invariants such as the Milnor number $\mu_f$, the perverse sheaf of vanishing cycles $\mathsf{P}_{U,f}$ and the category of matrix factorizations $\mathsf{MF}(U,f)$ are naturally defined on the Darboux stack, without ambiguity. The stack of non-degenerate flat quadratic bundles acts on the Darboux stack and our main theorem is the contractibility of the quotient stack when taking a further homotopy quotient identifying isotopic automorphisms. As a corollary we recover the gluing results for vanishing cycles by Brav--Bussi--Dupont--Joyce--Szendr\H oi. In a second part (to appear), we will apply this general mechanism to glue the motives of the locally defined categories of matrix factorizations $\mathsf{MF}(U,f)$ under the prescription of additional orientation data, thus answering positively conjectures by Kontsevich--Soibelman and Toda in motivic Donaldson--Thomas theory., Comment: 106 pages. Comments and feedback are of course very welcome

The interplay between T Tauri stars and their circumstellar disks, and how this impacts the onset of planet formation has yet to be established. We studied a seemingly old T Tauri star, PDS 111, and its disk. We analyzed optical, infrared, and sub-millimeter observations obtained with VLT/X-shooter, Mercator/HERMES, TESS, VLT/SPHERE, and ALMA, providing a new view on PDS 111 and its protoplanetary disk. The multi-epoch spectroscopy yields photospheric lines to classify the star, and emission lines to study variability in the hot inner disk and to determine the mass-accretion rate. The SPHERE and ALMA observations are used to characterize the dust distribution of the small and large grains, respectively. PDS 111 is a weak-line T Tauri star with spectral type G2, exhibits strong H$\alpha$ variability and with a low mass-accretion rate of $1-5\times10^{-10}$\,M$_{\odot}$\,yr$^{-1}$. We measured an age of the system of 15.9$^{+1.7}_{-3.7}$ Myr using pre-main sequence tracks. The SPHERE observations show a strongly flaring disk with an asymmetric substructure. The ALMA observations reveal a 30 au cavity in the dust continuum emission with a low contrast asymmetry in the South-West of the disk and a dust disk mass of 45.8\,$M_\oplus$. The $^{12}$CO radial extension is at least three times larger than that of the dust emission. Although the measured age is younger than suggested in literature, PDS 111 still seems relatively old; this provides insight into disk properties at an advanced stage of pre-main sequence evolution. The characteristics of this disk are very similar to its younger counterparts: strongly flaring, an average disk mass, a typical radial extent of the disk gas and dust, and the presence of common substructures. This suggests that disk evolution has not significantly changed the disk properties. These results show similarities with the "Peter Pan disks" around M-dwarfs., Comment: 23 pages, 23 figures, accepted by A&A (abstract shortened)

As public sector agencies rapidly introduce new AI tools in high-stakes domains like social services, it becomes critical to understand how decisions to adopt these tools are made in practice. We borrow from the anthropological practice to ``study up'' those in positions of power, and reorient our study of public sector AI around those who have the power and responsibility to make decisions about the role that AI tools will play in their agency. Through semi-structured interviews and design activities with 16 agency decision-makers, we examine how decisions about AI design and adoption are influenced by their interactions with and assumptions about other actors within these agencies (e.g., frontline workers and agency leaders), as well as those above (legal systems and contracted companies), and below (impacted communities). By centering these networks of power relations, our findings shed light on how infrastructural, legal, and social factors create barriers and disincentives to the involvement of a broader range of stakeholders in decisions about AI design and adoption. Agency decision-makers desired more practical support for stakeholder involvement around public sector AI to help overcome the knowledge and power differentials they perceived between them and other stakeholders (e.g., frontline workers and impacted community members). Building on these findings, we discuss implications for future research and policy around actualizing participatory AI approaches in public sector contexts.

As organizations face the challenges of processing exponentially growing data volumes, their reliance on analytics to unlock value from this data has intensified. However, the intricacies of big data, such as its extensive feature sets, pose significant challenges. A crucial step in leveraging this data for insightful analysis is an in-depth understanding of both the data and its domain. Yet, existing literature presents a fragmented picture of what comprises an effective understanding of data and domain, varying significantly in depth and focus. To address this research gap, we conduct a systematic literature review, aiming to delineate the dimensions of data understanding. We identify five dimensions: Foundations, Collection & Selection, Contextualization & Integration, Exploration & Discovery, and Insights. These dimensions collectively form a comprehensive framework for data understanding, providing guidance for organizations seeking meaningful insights from complex datasets. This study synthesizes the current state of knowledge and lays the groundwork for further exploration., Comment: Accepted at 32nd European Conference on Information Systems (2024)

We carried out very deep VLT/SPHERE imaging polarimetry of the nearby system Eps Eri based on 38.5 hours of integration time with a 600 - 900 nm broadband filter to search for polarized scattered light from a planet or from circumstellar dust using AO, coronagraphy, high precision differential polarimetry, and angular differential imaging. We have improved several data reduction and post-processing techniques and also developed new ones to further increase the sensitivity of SPHERE/ZIMPOL. The data provide unprecedented contrast limits, but no significant detection of a point source or an extended signal from circumstellar dust. For each observing epoch, we obtained a point source contrast for the polarized intensity between $2\cdot 10^{-8}$ and $4\cdot 10^{-8}$ at the expected separation of the planet Eps Eri b of 1'' near quadrature phase. The polarimetric contrast limits are about six to 50 times better than the intensity limits because polarimetric imaging is much more efficient in speckle suppression. Combining the entire 14-month data set to the search for a planet moving on a Keplerian orbit with the K-Stacker software further improves the contrast limits by a factor of about two, to about $8 \cdot 10^{-9}$ at 1''. This would allow the detection of a planet with a radius of about 2.5 Jupiter radii. The surface brightness contrast limits achieved for the polarized intensity from an extended scattering region are about 15 mag arcsec$^{-2}$ at 1'', or up to 3 mag arcsec$^{-2}$ deeper than previous limits. For Eps Eri, these limits exclude the presence of a narrow dust ring and they constrain the dust properties. This study shows that the polarimetric contrast limits for reflecting planets with SPHERE/ZIMPOL can be improved to a level $<10^{-8}$ simply by collecting more data over many nights and using the K-Stacker software., Comment: 26 pages, 20 figures, 5 tables. Accepted for publication in A&A

Context. Most ground-based planet search direct imaging campaigns use angular differential imaging, which distorts the signal from extended sources like protoplanetary disks. In the case PDS 70, a young system with two planets found within the cavity of a protoplanetary disk, obtaining a reliable image of both planets and disk is essential to understanding planet-disk interactions. Aims. Our goals are to reveal the true intensity of the planets and disk without self-subtraction effects for the first time, search for new giant planets beyond separations of 0.1" and to study the morphology of the disk shaped by two massive planets. Methods. We present YJHK-band imaging, polarimetry, and spatially resolved spectroscopy of PDS 70 using near-simultaneous reference star differential imaging, also known as star-hopping. We created a radiative transfer model of the system to match the near-infrared imaging and polarimetric data, along with sub-millimeter imaging from ALMA. Furthermore, we extracted the spectra of the planets and the disk and compared them. Results. We find that the disk is quite flared with a scale height of ~15% at the outer edge of the disk at ~90 au, similar to some disks in the literature. The gap inside of ~50 au is estimated to have ~1% of the dust density of the outer disk. The Northeast outer disk arc seen in previous observations is likely the outer lip of the flared disk. Abundance ratios of grains estimated by the modeling indicate a shallow grain-size index > -2.7, instead of the canonical -3.5. There is both vertical and radial segregation of grains. Planet c is well separated from the disk and has a spectrum similar to planet b, clearly redder than the disk spectra. Planet c is possibly associated with the sudden flaring of the disk starting at ~50 au. No new planets > 5 Mj were found., Comment: Accepted to A&A on April 11, 2024. 20 pages, 19 figures

This paper presents a study on performance optimization and resonant frequency modification of terahertz detectors by the use of hyper-hemispherical silicon superstrate lenses. The detectors are patch-TeraFETs, i.e., field-effect transistors with monolithically integrated patch antennas fabricated with a commercial 65-nm CMOS foundry process and designed for an operation frequency of 580 GHz. We demonstrate a strong improvement of the optical noise-equivalent power optical NEP, referenced against the total radiation power) reaching a value of 16 pW/Hz^(1/2). We show, furthermore, that the resonance frequency can be efficiently fine-tuned by the choice of the material and the thickness of a dielectric layer placed between the transistor and the superstrate lens. The resonance frequency can be shifted by more than 15 % of the center frequency (up to 100 GHz for the 580 GHz devices). The design of the on-chip optics can be employed for post-fabrication tailoring of the detector's resonance frequency to target specific spectral positions., Comment: 11 pages, 7 figures

Monolithically integrated, antenna-coupled field-effect transistors (TeraFETs) are rapid and sensitive detectors for the terahertz range (0.3-10~THz) that can operate at room temperature. We conducted experimental characterizations of a single patch-antenna coupled TeraFET optimized for 3.4~THz operation and its integration into an 8x8 multi-element detector configuration. In this configuration, the entire TeraFET array operates as a unified detector element, combining the output signals of all detector elements. Both detectors were realized using a mature commercial Si-CMOS 65-nm process node. Our experimental characterization employed single-mode Quantum-Cascade Lasers (QCLs) emitting at 2.85~THz and 3.4~THz. The 8x8 multi-element detector yields two major improvements for sensitive power detection experiments. First, the larger detector area simplifies alignment and enhances signal stability. Second, the reduced detector impedance enabled the implementation of a TeraFET+QCL system capable of providing a -3~dB modulation bandwidth up to 21~MHz, which is currently limited primarily by the chosen readout circuitry. Finally, we validate the system's performance by providing high resolution gas spectroscopy data for methanol vapor around 3.4~THz, where a detection limit of 1.6e-5 absorbance, or 2.6e11~molecules/cm^3 was estimated under optimal coupling conditions., Comment: accepted for publication in IEEE Transactions on Terahertz Science and Technology

We report on the modeling and experimental characterization of Si CMOS detectors of terahertz radiation based on antenna-coupled field-effect transistors (TeraFETs). The detectors are manufactured using TSMC's 65-nm technology. We apply two models -- the TSMC RF foundry model and our own ADS-HDM -- to simulate the Si CMOS TeraFET performance and compare their predictions with respective experimental data. Both models are implemented in the commercial circuit simulation software Keysight Advanced Design System (ADS). We find that the compact model TSMC RF is capable to predict the detector responsivity and its dependence on frequency and gate voltage with good accuracy up to the highest frequency of 1.2 THz covered in this study. This frequency is well beyond the tool's intended operation range for 5G communications and 110-GHz millimeter wave applications. We demonstrate that our self-developed physics-based ADS-HDM tool, which relies on an extended one-dimensional hydrodynamic transport model and can be adapted readily to other material technologies, has high predictive qualities comparable to those of the foundry model. We use the ADS-HDM to discuss the contribution of diffusive and plasmonic effects to the THz response of Si CMOS TeraFETs, finding that these effects, while becoming more significant with rising frequency, are never dominant. Finally, we estimate that the electrical NEP (perfect power coupling conditions) is on the order of 5 pW/$\sqrt{\rm{Hz}}$ at room-temperature.

Predictive models are often introduced to decision-making tasks under the rationale that they improve performance over an existing decision-making policy. However, it is challenging to compare predictive performance against an existing decision-making policy that is generally under-specified and dependent on unobservable factors. These sources of uncertainty are often addressed in practice by making strong assumptions about the data-generating mechanism. In this work, we propose a method to compare the predictive performance of decision policies under a variety of modern identification approaches from the causal inference and off-policy evaluation literatures (e.g., instrumental variable, marginal sensitivity model, proximal variable). Key to our method is the insight that there are regions of uncertainty that we can safely ignore in the policy comparison. We develop a practical approach for finite-sample estimation of regret intervals under no assumptions on the parametric form of the status quo policy. We verify our framework theoretically and via synthetic data experiments. We conclude with a real-world application using our framework to support a pre-deployment evaluation of a proposed modification to a healthcare enrollment policy., Comment: ICML 2024

In recent years, Large Language Models (LLMs) have shown remarkable performance in generating human-like text, proving to be a valuable asset across various applications. However, adapting these models to incorporate new, out-of-domain knowledge remains a challenge, particularly for facts and events that occur after the model's knowledge cutoff date. This paper investigates the effectiveness of Supervised Fine-Tuning (SFT) as a method for knowledge injection in LLMs, specifically focusing on the domain of recent sporting events. We compare different dataset generation strategies -- token-based and fact-based scaling -- to create training data that helps the model learn new information. Our experiments on GPT-4 demonstrate that while token-based scaling can lead to improvements in Q&A accuracy, it may not provide uniform coverage of new knowledge. Fact-based scaling, on the other hand, offers a more systematic approach to ensure even coverage across all facts. We present a novel dataset generation process that leads to more effective knowledge ingestion through SFT, and our results show considerable performance improvements in Q&A tasks related to out-of-domain knowledge. This study contributes to the understanding of domain adaptation for LLMs and highlights the potential of SFT in enhancing the factuality of LLM responses in specific knowledge domains., Comment: 16 pages; 7 figures. updated authors list

Eruptive stars are a class of young stellar objects that show an abrupt increase in their luminosity. These burst-like episodes are thought to dominate the stellar accretion process during the class 0/class I stage. We present an overview of a survey of seven episodically accreting protostars aimed at studying their potentially complex circumstellar surroundings. The observations were performed with the instrument SPHERE, mounted at the VLT. We observed the eruptive stars in $H$-band with the near-infrared imager IRDIS and used the polarimeter to extract the polarized light scattered from the stars' surroundings. We produced polarized light images for three FUor objects, Z CMa, V960 Mon, and FU Ori, and four EXor objects, XZ Tau, UZ Tau, NY Ori, and EX Lup. We calculated the intrinsic polarization fraction for all the observed stars. In all systems we registered scattered light from around the primary star. FU Ori and V960 Mon are surrounded by complex structures including spiral-like features. In Z CMa, we detected a point source 0.7 arcsec to the northeast of the primary. Based on the astrometric measurements from archival Keck/NIRC2 data, we find this source to be a third member of the system. Further, Z CMa displays an outflow extending thousands of au. Unlike the other EXor objects in our sample, XZ Tau shows bright, extended scattered light structures, also associated with an outflow on a scale of hundreds of au. The other EXors show relatively faint disk-like structures in the immediate vicinity of the coronagraph. Asymmetric arms were only found around FUor objects, while faint disks seem to predominantly occur around EXors. Importantly, for Z CMa the detection of the faint extended structure questions previous interpretations of the system's dynamic state. The streamer which was associated with a fly-by object turned out to be part of a huge outflow extending 6000 au., Comment: Accepted for publication in A&A. 14 pages, 7 figures, 2 tables

The sample of planet-forming disks observed by high-contrast imaging campaigns over the last decade is mature enough to enable the demographical analysis of individual star-forming regions. We present the full census of Taurus sources with VLT/SPHERE polarimetric images available. The whole sample sums up to 43 targets (of which 31 have not been previously published) corresponding to one-fifth of the Class II population in Taurus and about half of such objects that are observable. A large fraction of the sample is apparently made up of isolated faint disks (equally divided between small and large self-shadowed disks). Ambient signal is visible in about one-third of the sample. This probes the interaction with the environment and with companions or the outflow activity of the system. The central portion of the Taurus region almost exclusively hosts faint disks, while the periphery also hosts bright disks interacting with their surroundings. The few bright disks are found around apparently older stars. The overall picture is that the Taurus region is in an early evolutionary stage of planet formation. Yet, some objects are discussed individually, as in an intermediate or exceptional stage of the disk evolution. This census provides a first benchmark for the comparison of the disk populations in different star forming regions., Comment: Accepted by A&A

We present SPHERE/IRDIS H-band data for a sample of 23 stars in the Orion Star forming region observed within the DESTINYS (Disk Evolution Study Through Imaging of Nearby Young Stars) program. We use polarization differential imaging in order to detect scattered light from circumstellar dust. From the scattered light observations we characterize the disk orientation, radius and contrast. We analyse the disks in context of the stellar parameters and the environment of the Orion star-forming region. We use ancillary X-shooter spectroscopic observations to characterize the central stars in the systems. We furthermore use a combination of new and archival ALMA mm-continuum observations to characterize the dust masses present in the circumstellar disks. Within our sample we detect extended circumstellar disks in 10 of 23 systems. Of these, three are exceptionally extended (V351 Ori, V599 Ori and V1012 Ori) and show scattered light asymmetries which may indicate perturbations by embedded planets or (in the case of V599 Ori) by an outer stellar companion. Our high resolution imaging observations are also sensitive to close (sub)stellar companions and we detect 9 such objects in our sample of which 5 were previously unknown. We find in particular a possible sub-stellar companion (either a very low mass star or a high mass brown dwarf) 137 au from the star RY Ori. We find a strong anti-correlation between disk detection and multiplicity, with only 2 of our 10 disk detections located in stellar multiple systems. We also find a correlation between scattered light contrast and the millimetre flux suggesting that disks that have a high dust content are typically bright in near-infrared scattered light. Conversely we do not find significant correlations between scattered light contrast of the disks and the stellar mass or age., Comment: Accepted by A&A

We used VLT/SPHERE to observe 20 systems in the Cha I cloud in polarized scattered light in the near-infrared. We combined the scattered light observations with existing literature data on stellar properties and with archival ALMA continuum data to study trends with system age and dust mass. We also connected resolved near-infrared observations with the spectral energy distributions of the systems. In 13 of the 20 systems included in this study we detected resolved scattered light signals from circumstellar dust. For the CR Cha, CT Cha, CV Cha, SY Cha, SZ Cha, and VZ Cha systems we present the first detailed descriptions of the disks in scattered light. The observations found typically smooth or faint disks, often with little substructure, with the notable exceptions of SZ Cha, which shows an extended multiple-ringed disk, and WW Cha, which shows interaction with the cloud environment. New high S/N K- band observations of the HD 97048 system in our survey reveal a significant brightness asymmetry that may point to disk misalignment and subsequent shadowing of outer disk regions, possibly related to the suggested planet candidate in the disk. We resolve for the first time the stellar binary in the CS Cha system. Multiple wavelength observations of the disk around CS Cha have revealed that the system contains small, compact dust grains that may be strongly settled, consistent with numerical studies of circumbinary disks. We find in our sample that there is a strong anti-correlation between the presence of a (close) stellar companion and the detection of circumstellar material with five of our seven nondetections located in binary systems., Comment: Accepted by A&A

Public sector agencies are rapidly deploying AI systems to augment or automate critical decisions in real-world contexts like child welfare, criminal justice, and public health. A growing body of work documents how these AI systems often fail to improve services in practice. These failures can often be traced to decisions made during the early stages of AI ideation and design, such as problem formulation. However, today, we lack systematic processes to support effective, early-stage decision-making about whether and under what conditions to move forward with a proposed AI project. To understand how to scaffold such processes in real-world settings, we worked with public sector agency leaders, AI developers, frontline workers, and community advocates across four public sector agencies and three community advocacy groups in the United States. Through an iterative co-design process, we created the Situate AI Guidebook: a structured process centered around a set of deliberation questions to scaffold conversations around (1) goals and intended use or a proposed AI system, (2) societal and legal considerations, (3) data and modeling constraints, and (4) organizational governance factors. We discuss how the guidebook's design is informed by participants' challenges, needs, and desires for improved deliberation processes. We further elaborate on implications for designing responsible AI toolkits in collaboration with public sector agency stakeholders and opportunities for future work to expand upon the guidebook. This design approach can be more broadly adopted to support the co-creation of responsible AI toolkits that scaffold key decision-making processes surrounding the use of AI in the public sector and beyond.

AI tools are increasingly deployed in community contexts. However, datasets used to evaluate AI are typically created by developers and annotators outside a given community, which can yield misleading conclusions about AI performance. How might we empower communities to drive the intentional design and curation of evaluation datasets for AI that impacts them? We investigate this question on Wikipedia, an online community with multiple AI-based content moderation tools deployed. We introduce Wikibench, a system that enables communities to collaboratively curate AI evaluation datasets, while navigating ambiguities and differences in perspective through discussion. A field study on Wikipedia shows that datasets curated using Wikibench can effectively capture community consensus, disagreement, and uncertainty. Furthermore, study participants used Wikibench to shape the overall data curation process, including refining label definitions, determining data inclusion criteria, and authoring data statements. Based on our findings, we propose future directions for systems that support community-driven data curation.

There are two common ways in which developers are incorporating proprietary and domain-specific data when building applications of Large Language Models (LLMs): Retrieval-Augmented Generation (RAG) and Fine-Tuning. RAG augments the prompt with the external data, while fine-Tuning incorporates the additional knowledge into the model itself. However, the pros and cons of both approaches are not well understood. In this paper, we propose a pipeline for fine-tuning and RAG, and present the tradeoffs of both for multiple popular LLMs, including Llama2-13B, GPT-3.5, and GPT-4. Our pipeline consists of multiple stages, including extracting information from PDFs, generating questions and answers, using them for fine-tuning, and leveraging GPT-4 for evaluating the results. We propose metrics to assess the performance of different stages of the RAG and fine-Tuning pipeline. We conduct an in-depth study on an agricultural dataset. Agriculture as an industry has not seen much penetration of AI, and we study a potentially disruptive application - what if we could provide location-specific insights to a farmer? Our results show the effectiveness of our dataset generation pipeline in capturing geographic-specific knowledge, and the quantitative and qualitative benefits of RAG and fine-tuning. We see an accuracy increase of over 6 p.p. when fine-tuning the model and this is cumulative with RAG, which increases accuracy by 5 p.p. further. In one particular experiment, we also demonstrate that the fine-tuned model leverages information from across geographies to answer specific questions, increasing answer similarity from 47% to 72%. Overall, the results point to how systems built using LLMs can be adapted to respond and incorporate knowledge across a dimension that is critical for a specific industry, paving the way for further applications of LLMs in other industrial domains.

The constantly increasing capabilities of artificial intelligence (AI) open new possibilities for human-AI collaboration. One promising approach to leverage existing complementary capabilities is allowing humans to delegate individual instances to the AI. However, enabling humans to delegate instances effectively requires them to assess both their own and the AI's capabilities in the context of the given task. In this work, we explore the effects of providing contextual information on human decisions to delegate instances to an AI. We find that providing participants with contextual information significantly improves the human-AI team performance. Additionally, we show that the delegation behavior changes significantly when participants receive varying types of contextual information. Overall, this research advances the understanding of human-AI interaction in human delegation and provides actionable insights for designing more effective collaborative systems.

We define Hochschild cohomology of the second kind for differential graded (dg) or curved algebras as a derived functor in a compactly generated derived category of the second kind, and show that it is invariant under Morita equivalence of the second kind. A bimodule version of Koszul duality is constructed and used to show that Hochschild cohomology of the second kind is preserved under (nonconilpotent) Koszul duality. Hochschild cohomology of the second kind of an algebra often computes the ordinary Hochschild cohomology of geometrically meaningful dg categories. Examples include the category of infinity local systems on a topological space, the bounded derived category of a complex algebraic manifold and the category of matrix factorizations., Comment: 23 pages

Polarisation is a powerful remote-sensing tool to study the nature of particles scattering the starlight. It is widely used to characterise interplanetary dust particles in the Solar System and increasingly employed to investigate extrasolar dust in debris discs' systems. We aim to measure the scattering properties of the dust from the debris ring around HD 181327 at near-infrared wavelengths. We obtained high-contrast polarimetric images of HD 181327 in the H band with the SPHERE / IRDIS instrument on the Very Large Telescope (ESO). We complemented them with archival data from HST / NICMOS in the F110W filter reprocessed in the context of the Archival Legacy Investigations of Circumstellar Environments (ALICE) project. We developed a combined forward-modelling framework to simultaneously retrieve the scattering phase function in polarisation and intensity. We detected the debris disc around HD 181327 in polarised light and total intensity. We measured the scattering phase function and the degree of linear polarisation of the dust at 1.6 micron in the birth ring. The maximum polarisation is 23.6% +/- 2.6% and occurs between a scattering angle of 70 deg and 82 deg. We show that compact spherical particles made of a highly refractive and relatively absorbing material in a differential power-law size distribution of exponent $-3.5$ can simultaneously reproduce the polarimetric and total intensity scattering properties of the dust. This type of material cannot be obtained with a mixture of silicates, amorphous carbon, water ice, and porosity, and requires a more refracting component such as iron-bearing minerals. We reveal a striking analogy between the near-infrared polarisation of comets and that of HD 181327. The methodology developed here combining VLT/SPHERE and HST/NICMOS may be applicable in the future to combine the polarimetric capabilities of SPHERE with the sensitivity of JWST., Comment: 21 pages, 22 figures, accepted in Astronomy & Astrophysics

Field-effect transistors coupled to integrated antennas (TeraFETs) are photodetectors being actively developed for the THz frequency range ($\sim$ 100 GHz - 10 THz). Among them, Graphene TeraFETs (G-TeraFETs) have demonstrated distinctive photoresponse features compared to those made from elementary semiconductors. For instance, previous studies have shown that G-TeraFETs exhibit a THz response that comprises two components: the resistive self-mixing (RSM) and photothermoelectric effect (PTE). The RSM and PTE arise from carrier density oscillations and carrier heating, respectively. In this work, we confirm that the photoresponse can be considered a combination of RSM and PTE, with PTE being the dominant rectification mechanism at higher frequencies. For our CVD G-TeraFETs with asymmetric antenna coupling, the PTE response dominates over the RSM at frequencies above 100 GHz. We find that relative contribution of RSM and PTE to the photoresponse is strongly frequency dependent. Electromagnetic wave simulations show that this behavior is due to the relative change in the total dissipated power between the gated and ungated channel regions of the G-TeraFET as the frequency increases. The simulations also indicate that the channel length over which the PTE contributes to the photoresponse below the gate electrode is approximately the same as the electronic cooling length. Finally, we identify a PTE contribution that can be attributed to the contact doping effect in graphene close to the metal contacts. Our detectors achieve a minimum optical noise-equivalent power of 101 (114) pW/$\sqrt{Hz}$ for asymmetric (symmetric) THz antenna coupling conditions at 400 GHz. This work demonstrates how the PTE response can be used to optimize the THz responsivity of G-TeraFETs., Comment: new title and abstract, 32 pages and 14 figures

Context. The study of protoplanetary disks is fundamental to understand their evolution and interaction with the surrounding environment, and to constrain planet formation mechanisms. Aims. We aim at characterising the young binary system HD 34700 A, which shows a wealth of structures. Methods. Taking advantage of the high-contrast imaging instruments SPHERE at the VLT, LMIRCam at the LBT, and of ALMA observations, we analyse this system at multiple wavelengths. We study the rings and spiral arms morphology and the scattering properties of the dust. We discuss the possible causes of all the observed features. Results. We detect for the first time, in the H${\alpha}$ band, a ring extending from $\sim$65 au to ${\sim}$120 au, inside the ring already known from recent studies. These two have different physical and geometrical properties. Based on the scattering properties, the outer ring may consist of grains of typical size $a_{out} > 4 {\mu}m$, while the inner ring of smaller grains ($a_{in} <= 0.4 {\mu m}$). Two extended logarithmic spiral arms stem from opposite sides of the disk. The outer ring appears as a spiral arm itself, with a variable radial distance from the centre and extended substructures. ALMA data confirm the presence of a millimetric dust substructure centred just outside the outer ring, and detect misaligned gas rotation patterns for HD 34700 A and B. Conclusions. The complexity of HD 34700 A, revealed by the variety of observed features, suggests the existence of one or more disk-shaping physical mechanisms. Possible scenarios, compatible with our findings, involve the presence inside the disk of a yet undetected planet of several Jupiter masses and the system interaction with the surroundings by means of gas cloudlet capture or flybys. Further observations with JWST/MIRI or ALMA (gas kinematics) could shed more light on these., Comment: Accepted for publication on A&A. 14 + 5 pages, 9 + 7 figures (text + appendix)

While there have been nearly two dozen of spiral arms detected from planet-forming disks in near-infrared scattered light, none of their substellar drivers have been confirmed. By observing spiral systems in at least two epochs spanning multiple years, and measuring the motion of the spirals, we can distinguish the cause of the spirals, and locate the orbits of the driving planets if they trigger the spirals. Upon a recent validation of this approach using the co-motion between a stellar companion and a spiral, we obtained a second epoch observation for the spiral system in the disk of V1247 Ori in the $H$-band polarized scattered light using VLT/SPHERE/IRDIS. Combining our observations with archival IRDIS data, we established a $4.8$ yr timeline to constrain the V1247 Ori spiral motion. We obtained a pattern speed of $0.40^{\circ} \pm 0.09^{\circ}$ yr$^{-1}$ for the north-east spiral. This corresponds to an orbital period of $900\pm200$ yr, and thus the semi-major axis of the hidden planetary driver is $118\pm19$ au for a 2.0 $\pm$ 0.1 M$_\odot$ central star. The location agrees with the gap in ALMA dust continuum observations, providing joint support for the existence of a companion driving the scattered-light spirals while carving a millimeter gap. With an angular separation of 0.29" $\pm$ 0.05", this hidden companion is an ideal target for JWST imaging., Comment: Accepted for publication in Astronomy and Astrophysics; 6 pages, 5 figures

The Coronagraph Instrument of the Nancy Grace Roman Space Telescope (Roman Coronagraph) will be capable of both total intensity and polarization measurements of circumstellar disks. The polarimetric performance is impacted by polarization effects introduced by all mirrors before the Wollaston prisms. In this paper, we aim to characterize these effects for the Roman Coronagraph in bands 1 and 4 using the FALCO and PROPER packages. We simulate the effect of polarization aberrations that impact the polarimetric contrast and the instrumental polarization effects to study the polarimetric accuracy. We include spacecraft rolls, but leave out systematic camera noise. We find that polarimetric differential imaging (PDI) improves the contrast by a factor of six. The PDI contrast of $\sim 8 \times 10^{-11}$ is limited by polarized speckles from instrumental polarization effects and polarization aberrations. By injecting polarized companions with at various contrast levels and demodulating their polarimetric signal, we recover their source Stokes vector within 2%., Comment: 16 pages, 16 figures, SPIE Optics + Photonics - Techniques and Instrumentation for Detection of Exoplanets XI

A growing body of research has explored how to support humans in making better use of AI-based decision support, including via training and onboarding. Existing research has focused on decision-making tasks where it is possible to evaluate "appropriate reliance" by comparing each decision against a ground truth label that cleanly maps to both the AI's predictive target and the human decision-maker's goals. However, this assumption does not hold in many real-world settings where AI tools are deployed today (e.g., social work, criminal justice, and healthcare). In this paper, we introduce a process-oriented notion of appropriate reliance called critical use that centers the human's ability to situate AI predictions against knowledge that is uniquely available to them but unavailable to the AI model. To explore how training can support critical use, we conduct a randomized online experiment in a complex social decision-making setting: child maltreatment screening. We find that, by providing participants with accelerated, low-stakes opportunities to practice AI-assisted decision-making in this setting, novices came to exhibit patterns of disagreement with AI that resemble those of experienced workers. A qualitative examination of participants' explanations for their AI-assisted decisions revealed that they drew upon qualitative case narratives, to which the AI model did not have access, to learn when (not) to rely on AI predictions. Our findings open new questions for the study and design of training for real-world AI-assisted decision-making.

(Abridged) Context. To directly image rocky exoplanets in reflected (polarized) light, future space- and ground-based high-contrast imagers and telescopes aim to reach extreme contrasts at close separations from the star. However, the achievable contrast will be limited by reflection-induced polarization aberrations. While polarization aberrations can be modeled numerically, such computations provide little insight into the full range of effects, their origin and characteristics, and possible ways to mitigate them. Aims. We aim to understand polarization aberrations produced by reflection off flat metallic mirrors at the fundamental level. Methods. We used polarization ray tracing to numerically compute polarization aberrations and interpret the results in terms of the polarization-dependent spatial and angular Goos-H\"anchen and Imbert-Federov shifts of the beam of light as described with closed-form mathematical expressions in the physics literature. Results. We find that all four beam shifts are fully reproduced by polarization ray tracing and study the origin, characteristics, sizes, and directions of the shifts. Of the four beam shifts, only the spatial Goos-H\"anchen and Imbert-Federov shifts are relevant for high-contrast imagers and telescopes because these shifts are visible in the focal plane and create a polarization structure in the PSF that reduces the performance of coronagraphs and the polarimetric speckle suppression close to the star. Conclusions. The beam shifts in an optical system can be mitigated by keeping the f-numbers large and angles of incidence small. Most importantly, mirror coatings should not be optimized for maximum reflectivity, but should be designed to have a retardance close to 180{\deg}. The insights from our study can be applied to improve the performance of current and future high-contrast imagers, especially those in space and on the ELTs., Comment: 19 pages, 13 figures, 1 table, final version as published in Astronomy & Astrophysics, 677, A150 (2023)

An emerging body of research indicates that ineffective cross-functional collaboration -- the interdisciplinary work done by industry practitioners across roles -- represents a major barrier to addressing issues of fairness in AI design and development. In this research, we sought to better understand practitioners' current practices and tactics to enact cross-functional collaboration for AI fairness, in order to identify opportunities to support more effective collaboration. We conducted a series of interviews and design workshops with 23 industry practitioners spanning various roles from 17 companies. We found that practitioners engaged in bridging work to overcome frictions in understanding, contextualization, and evaluation around AI fairness across roles. In addition, in organizational contexts with a lack of resources and incentives for fairness work, practitioners often piggybacked on existing requirements (e.g., for privacy assessments) and AI development norms (e.g., the use of quantitative evaluation metrics), although they worry that these tactics may be fundamentally compromised. Finally, we draw attention to the invisible labor that practitioners take on as part of this bridging and piggybacking work to enact interdisciplinary collaboration for fairness. We close by discussing opportunities for both FAccT researchers and AI practitioners to better support cross-functional collaboration for fairness in the design and development of AI systems., Comment: In Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency (FAccT '23)