Search for: All records

eXplainable Artificial Intelligence (XAI) has garnered significant attention for enhancing transparency and trust in machine learning models. However, the scopes of most existing explanation techniques focus either on offering a holistic view of the explainee model (global explanation) or on individual instances (local explanation), while the middle ground, i.e., cohort-based explanation, is less explored. Cohort explanations offer insights into the explainee's behavior on a specific group or cohort of instances, enabling a deeper understanding of model decisions within a defined context. In this paper, we discuss the unique challenges and opportunities associated with measuring cohort explanations, define their desired properties, and create a generalized framework for generating cohort explanations based on supervised clustering. 

Abstract In this paper, we explore the benefits that digitalization and Big Data analytics can bring to animal health, emphasizing the need to advance toward precision veterinary epidemiology. This concept takes advantage of multilevel animal health–related data to better understand disease dynamics in a population and design more cost-effective interventions, particularly focusing on livestock health. However, to translate this concept into practice, critical advancements and changes are needed in how we collect, standardize, integrate, share, and use data. Fostering interdisciplinary teams that integrate epidemiologists, veterinarians, and other domain experts with computer scientists, engineers, and data scientists is essential to implement this approach and better address complex animal health issues. 

Abstract Periodic variability in active galactic nuclei (AGNs) is a promising method for studying subparsec supermassive black hole binaries (SMBHBs), which are a challenging detection target. While extensive searches have been made in the optical, X-ray, and gamma-ray bands, systematic infrared (IR) studies remain limited. Using data from the Wide-field Infrared Survey Explorer (WISE), which provides unique decade-long mid-IR light curves with a six-month cadence, we have conducted the first systematic search for SMBHB candidates based on IR periodicity. Analyzing a parent sample of 48,932 objects selected from about half a million AGNs, we have identified 28 candidate periodic AGNs with periods ranging from 1268 to 2437 days (in the observer frame), by fitting their WISE light curves with sinusoidal functions. However, our mock simulation of the parent sample indicates that stochastic variability can actually produce a similar number of periodic sources, underscoring the difficulty in robustly identifying real periodic signals with WISE light curves, given their current sampling. Notably, we find no overlap between our sample and optical periodic sources, which can be explained by a distinct preference for certain periods due to selection bias. By combining archived data from different surveys, we have identified a candidate exhibiting periodic behavior in both the optical and IR bands, a phenomenon that warrants further validation through observational tests. Our results highlight the potential of IR time-domain surveys, including future missions such as the Nancy Grace Roman Space Telescope, for identifying periodic AGNs, but complementary tests are still needed to determine their physical origins, such as SMBHBs. 

The instruction hierarchy, which establishes a priority order from system messages to user messages, conversation history, and tool outputs, is essential for ensuring consistent and safe behavior in language models (LMs). Despite its importance, this topic receives limited attention, and there is a lack of comprehensive benchmarks for evaluating models’ ability to follow the instruction hierarchy. We bridge this gap by introducing IHEval, a novel benchmark comprising 3,538 examples across nine tasks, covering cases where instructions in different priorities either align or conflict. Our evaluation of popular LMs highlights their struggle to recognize instruction priorities. All evaluated models experience a sharp performance decline when facing conflicting instructions, compared to their original instruction-following performance. Moreover, the most competitive open-source model only achieves 48% accuracy in resolving such conflicts. Our results underscore the need for targeted optimization in the future development of LMs. 

Abstract Designing CRISPR single guide RNA (sgRNA) libraries targeting entire kingdoms of life will significantly advance genetic research in diverse and underexplored taxa. Current sgRNA design tools are often species-specific and fail to scale to large, phylogenetically diverse datasets, limiting their applicability to comparative genomics, evolutionary studies, and biotechnology. Here, we present ALLEGRO, a combinatorial optimization algorithm able to design minimal, yet highly effective sgRNA libraries targeting thousands of species. Leveraging integer linear programming, ALLEGRO identified compact sgRNA sets simultaneously targeting several genes of interest for over 2,000 species across the fungal kingdom. We experimentally validated the sgRNAs designed by ALLEGRO inKluyveromyces marxianus, Komagataella phaffii, andYarrowia lipolytica. In addition, we adopted a generalized Cas9-Ribonucleoprotein delivery system coupled with protoplast transformation to extend ALLEGRO’s sgRNA libraries to other untested fungal genomes, such asRhodotorula araucariae. Our experimental results, along with cross-validation, show that ALLEGRO enables efficient CRISPR genome editing, supporting the development of universal sgRNA libraries applicable to entire taxonomic groups. 

Abstract The search for dual supermassive black holes (SMBHs) is of immense interest in modern astrophysics. Galaxy mergers may fuel and produce SMBH pairs. Actively accreting SMBH pairs are observed as dual quasars, which are vital probes of SMBH growth. Dual quasars at cosmic noon are not well characterized. Gaia observations have enabled a novel technique to identify dual quasars at kiloparsec scales based on the small jitters of the light centroid as the two quasars vary stochastically. We present the first detailed study of az= 2.17, 0 $\stackrel{\text{″}}{.}$46, 3.8 kpc separation dual quasar, J0749+2255, using JWST/NIRSpec integral field unit spectroscopy. Identified by Gaia, J0749+2255 is one of the most distant small-separation dual quasars known. We detect the faint ionized gas of the host galaxy, traced by the narrow Hαemission. Line ratios indicate ionization from the two quasars and from intense star formation. Spectral analysis of the two quasars suggests that they have similar black hole properties, hinting at the possible synchronized accretion activity or lensed quasar images. Surprisingly, the ionized gas kinematics suggest a rotating disk rather than the disturbed system expected in a major gas-rich galaxy merger. Numerical simulations show that this is a plausible outcome of a major gas-rich galaxy merger several tens of Myr before coalescence. Whether J0749+2255 reflects an interesting phase of dual quasar evolution or is a lensed quasar remains unclear. Thus, this study underscores the challenges in definitively distinguishing between dual and lensed quasars, with observations supporting either scenario.