學術演講及專題研討

Modern observational surveys in optical and X-rays have continuously uncovered increasingly short-lived and faint transient phenomena, demanding advances in theory to keep pace. I will discuss recent theoretical progress on three types of events: luminous red novae associated with stellar mergers, Type II supernovae with photometry on sub-day timescales, and fast X-ray transients discovered by Einstein Probe. I will first investigate what can be gleaned about the power source of luminous red novae from their complex spectral evolution. Next, I will present a study of the puzzling red-to-blue color evolution during the early rising phase of the nearby SNe 2023ixf and 2024ggi, which indicates the presence of a steep density gradient in the immediate vicinity of their red supergiant SN progenitors. Finally, I will consider the shock breakout candidate EP 260321a recently reported by Einstein Probe and what it may reveal about the nature of Type Ib/c supernovae and their central engines.

Over the past few decades, statistical analyses of the cosmic microwave background (CMB) have established the standard model of cosmology with remarkable precision. A natural next step is to extend these statistical techniques to the integrated molecular and atomic line emission from galaxies across cosmic time. Line intensity mapping (LIM) has emerged as a powerful observational approach that enables us to study the aggregate emission from unresolved galaxies, opening a new window on galaxy formation and evolution. In this seminar, I will introduce the basic physics of molecular and atomic line emission and discuss how current and upcoming LIM experiments aim to map these signals over large cosmological volumes. I will then show how cross-correlations between LIM observations, galaxy surveys, and the 21cm signal can reveal the physical properties of galaxies and the evolution of large-scale structure. I will also discuss why LIM measurements are important for precision CMB studies, particularly on small angular scales, where molecular line emission (such as CO) can act as a significant foreground for measurements of CMB secondary anisotropies. Finally, I will present our recent detection of CO(3–2) emission from galaxies at (z~0.5), obtained by stacking Sloan Digital Sky Survey (SDSS) galaxies on CO maps reconstructed from Planck data, and discuss what this measurement demonstrates about the potential of cosmological surveys to address astrophysical questions.

Molecular clouds are cradles of star-formation, yet black holes are the ultimate end products of stellar evolution. Molecular clouds are diffuse media spanning a wide range of parameters, while astrophysical black holes are extreme compact objects specified only by their mass and spin. Most studies of molecular clouds would not include any black holes in the consideration, and most studies of black holes of stellar origins are not associated with any molecular clouds. This work will explore what black holes "would do" when residing in and traversing across a molecular cloud and how their actions would impact on the structures and determine fate of their host molecular clouds. We will first show that the encounter of black hole and molecular clouds are not rare occasions but more frequent than we would expect if we take the face value of abundance of molecular clouds and the populations of stellar remnant black holes in our Galaxy. We will discuss the accretion of gas by black holes in molecular and their observational signatures, and the impacts on a molecular cloud when it is embedded with one or multiple black holes.

A presentation of IRAM, the Institute for Radio Astronomy in the Millimeter range in Grenoble France will be made. After a brief description of IRAM's mission, organization, and key figures, the presentation will cover IRAM's facilities, including its laboratories and observatories. In the following, a more detailed overview of the status of the IRAM observatories—the Northern Extended Millimeter Array (NOEMA) and the IRAM 30-meter telescope—will be given. Recent upgrades and their scientific impact will be described. Some scientific highlights and ongoing technical development for future upgrades will be presented.

Interaction with the surrounding media provides an important way to learn about the progenitors of supernovae (SNe). A fraction of core-collapse SNe show signs of strong interaction with the circumstellar material (CSM). Rising evidence points to asymmetry and sustained growth (over decades) of the CSM prior to core-collapse. Using a combination of the stellar evolution code MESA, a 3D hydrodynamics code, and a combination of radiation hydrodynamics and radiative transfer methods, I will demonstrate how SNe in binary systems can show signs of strong interaction, and may appear different when viewed from different directions. In the next part of my talk, I will venture ahead to the remnant phase and discuss how interaction in remnants of thermonuclear supernovae (Type-Ia SNRs) reveals the mass of the white dwarf (WD) progenitor. The mass of the WD in turn holds clue to the explosion mechanism, which is still poorly constrained for observed Type-Ia SNe. I will also demonstrate a technique to analyze X-ray and optical observations of SNRs. I will present application of the method to two well-known examples, unveiling whether their progenitor reached the Chandrasekhar mass at explosion.

The current generation of line intensity mapping (LIM) surveys face two primary observational limitations: sensitivity, and systematic errors from instrumental or astrophysical origins. Both of these limitations can be mitigated by combining LIM data with data from external galaxy surveys, which add priors about where signal should be located and are subject to systematic errors independent from those of the LIM survey. I will discuss a robust, simple statistical technique for performing this combination – stacking the LIM data on the positions of the survey galaxies. I will show stacking results from COMAP (the CO Mapping Array Project), a currently-observing LIM experiment targeting dense molecular gas in galaxies via their CO(1-0) emission. I will compare the output from COMAP stacks on quasars from eBOSS and DESI to modelling of the stacked emission and explore the implications of a mild tension between the models and the COMAP-eBOSS/DESI stacked upper limit. This simple method has the potential to illuminate the relationship between line-emitting gas and the large-scale environments of selected galaxy types, providing information beyond the scope of LIM auto-correlation studies.

More than 6000 exoplanets have been discovered, with their masses, radii, equilibrium temperatures, and atmospheres revealing a remarkable diversity in planetary properties and compositions. Within this landscape, sub-Saturn gas giant bridge sub-Neptunes and Jupiter-mass giants, forming a key population for understanding planetary interiors, evolution, and formation pathways. In this talk, I will review the fundamental theory of Solar System and exoplanet interiors, highlighting the mechanisms that shape the interior structure of sub-Saturns. I will address how observations and modeling are combined to constrain their interior compositions and structures. In particular, I will discuss results from studies based on mass–radius data, and show how additional constraints - such as age, transmission and emission spectra, and Love numbers - are emerging from current and upcoming observations with JWST.

To first approximation, the structure of large scale astronomical systems is governed by gravity, which is described by time-reversible ordinary differential equations. I present novel time-symmetric numerical methods for simulations of structure formation which respect this time-symmetry and show they can be significantly more accurate than non-reversible numerical methods in studying gravitational dynamics. One of these methods is being implemented as TRACE in the REBOUND software package, and we find for interesting planetary dynamics problems it offers a speed advantage of an order magnitude or more compared to other codes. Next, I focus on the effects of stellar flybys on Earth's climate history. Stellar flybys can have notable effects on the solar system's long-term dynamical evolution, injection of Oort cloud comets into the solar system, properties of trans-Neptunian objects, and more. Using a state-of-the-art solar system model, including a lunar contribution and the J2 solar quadrupole, and random stellar parameters, we find no influence of passing stars on paleoclimate reconstructions over the past 56 Myr, in contrast to recent results from other authors who did not include some of these physical effects. Finally, I tackle the problem of the stability of the Solar System. Although great progress has been made in the last decades towards an understanding of chaos and stability of the Solar System, I show that some studies are affected by numerical artifacts, which causes artificial Solar System chaos and instability. The physical mechanism behind Mercury's orbital instability has been traditionally described by a diffusive process in a secular frequency, but our current work shows a sub-diffusive process fits simulated data better. An explanation for this sub-diffusion remains elusive.

Our current understanding of the universe relies on the ΛCDM model that includes, among other components, a cosmological constant dark energy (Λ) and cold dark matter (CDM). The large-scale structure of the late universe contains a wealth of information that can be used to test this concordance model, and the second data release of Baryon Acoustic Oscillations from the the Dark Energy Spectroscopic Instrument (DESI) already places impressive constraints on cosmology. In particular, the w0waCDM parametrization of dark energy shows a slight preference for an evolving dark energy component, with an effective phantom behaviour at higher redshifts. If this phenomenological result persists, it would signal a major shift in our understanding, and require a new theoretical description of the universe. However, since this 2-parameter extension of ΛCDM might be prone to bias, or insensitive to certain behaviours, it is important to test the results using more flexible, model-independent reconstructions. I will focus on recent work within DESI where we do this by implementing crossing statistic reconstruction, binning and Gaussian process regression. What we find (SPOILER ALERT!) is a preference for an expansion history consistent with the trends seen in the Key Paper result (arXiv:2503.14738). And, if you are new to cosmology, don't let the jargon in the abstract scare you, I promise it's for official purposes only.

How disc galaxies form and evolve remains a fundamental question in astrophysics. The Milky Way provides a unique opportunity to address this, as we can measure precise positions, motions, and ages of individual stars. In particular, stellar kinematics offer a powerful way to uncover both current dynamical processes and the imprints of past evolution. In the first part of this work, by combining Gaia DR3 with classical Cepheids, we reveal the kinematic diversity of multiple spiral arms. The Perseus and Outer arms show opposite correlations between radial and azimuthal velocities. Comparison with simulations suggests that these two arms are in distinct evolutionary phases. These findings challenge the classical view of static spiral arms and instead support a dynamic, transient picture of spiral structure in the Milky Way. In the second part, we aim to trace how the Milky Way disc formed over time. Combining an action-based distribution function fitting with APOGEE–Gaia red giant stars with reliable age estimates allows us to reveal age-dependent trends in disc parameters. We identify the transition between the thick and thin disc populations, likely triggered by the Gaia-Sausage-Enceladus (GSE) merger. Together, these results show how stellar kinematics trace both the formation history and ongoing evolution of the Milky Way, and highlight the potential of future large-scale astrometric and spectroscopic surveys to further resolve its dynamics.

Recent ALMA observations have unveiled a wealth of fine structures in star-forming regions, such as protostellar outflows and jets, asymmetric gas accretion, and rotationally supported circumstellar disks. In addition, ring and gap substructures—often interpreted as signatures of planet formation—have been identified even during the active gas accretion phase. In this study, we carried out magnetohydrodynamic simulations of star formation starting from a prestellar core, and followed the subsequent formation and evolution of a protostar and its surrounding disk up to the Class II stage. Our simulations successfully reproduce the development of outflows and jets, as well as circumstellar disks exhibiting complex morphologies, envelopes, and asymmetric accretion flows. By comparing these results with ALMA observations, we gain insight into the physical processes governing the earliest stages of star and disk formation. In this talk, I will present recent progress in our understanding of early star formation based on comparisons between numerical simulations and observations.

Recent observations have revealed that circumbinary disks that are misaligned to the binary orbit could be common in the universe. Dissipation in the disk causes it to move either towards coplanar alignment or polar alignment. Since planets form inside disks, circumbinary planets may also form misaligned to the binary orbit. We explore the orbital evolution of the binary with the misaligned circumbinary disk. The disk-binary interaction leads to the evolution of the binary’s orbit, which could either shrink over time due to the loss of its angular momentum via resonances or expand due to the acquirement of specific angular momentum via accretion. We explore diverse parameter domains of circumbinary systems with polar and retrograde disks. Our results have implications in a variety of astrophysical objects, including the orbital evolution of stellar binaries and the evolution of both stellar and supermassive black hole binaries.

Very long baseline interferometry (VLBI) observations show that relativistic jets, such as the one in M87, exhibit a limb-brightened, double-edged structure that has been challenging to reproduce in theoretical models. In this talk, I present a radiative-transfer framework that combines anisotropic electron distributions with “slow-light” treatment, which considers the time-evolution of plasmas during light ray propagation. Motivated by recent particle-in-cell simulations, we assume that nonthermal electron velocities are preferentially aligned with magnetic field lines. Implementing this prescription in general relativistic magneto-hydrodynamic (GRMHD) models, we produce images across multiple frequencies and spatial scales. We find that synchrotron emission is naturally concentrated along helical magnetic fields, producing limb-brightened morphologies from tens of microarcseconds to hundreds of milliarcseconds in the M87 jet. When slow-light radiative transfer is included, loop-like features of the jet are stretched and smoothed, yielding a cone-like, double-edged jet morphology consistent with observations. We further show that jet appearance and variability depend sensitively on black hole spin, offering testable predictions for horizon-scale imaging with next-generation instruments such as EHT, ngEHT, and BHEX.

In 2019, the Event Horizon Telescope (EHT) collaboration released the first image of a supermassive black hole: the one residing at the center of the galaxy M87. Since that milestone, the EHT has established a rich legacy that includes the first images of Sagittarius A*, the black hole at the center of the Milky Way; the first images of black hole environments in polarized light—both linear and circular—revealing dynamically important magnetic fields near the event horizon and favoring magnetically arrested disk models; and the first strong evidence for intrinsic time variability in both total intensity and magnetic field structure. In this colloquium, I will review these key scientific results and discuss the evolution of the EHT array that made them possible. I will conclude by describing ongoing developments and future efforts aimed at obtaining the first time-resolved “movie” of a black hole, probing horizon-scale dynamics on timescales comparable to the characteristic timescales of the system.

Ultracold quantum gases play a pivotal role in many-body physics, quantum sensing, and quantum simulation. Over time, methods such as evaporative cooling in bulk ensembles and precision laser cooling techniques have been employed to achieve quantum degeneracy in atomic gases. The pursuit of a simpler, faster way to form quantum gases thus holds significant promise for advancing the field. In this talk, I will report on our creation of a quantum gas by cooling individual rubidium atoms pinned in a three-dimensional optical lattice using electromagnetically induced transparency and adiabatic expansion. After just 10 milliseconds of cooling, we verify the phase transition from a thermal to a quantum gas by adiabatically transferring the atoms into optical dipole traps. We observe the collapse of atoms in three-dimensional traps, a distinctive hallmark of a quantum gas with negative scattering length. Our results introduce a versatile and fast approach to achieving quantum degenerate gases with minimal time and resource requirements.

The fate of the most massive stars remains one of the central questions in stellar evolution and time-domain astronomy. Pair-Instability Supernovae (PISNe) and Pulsational PISNe (PPISNe), the predicted endpoints of stars above roughly 95 solar masses, offer a rare opportunity to test our understanding of stellar physics and to glimpse the processes that shaped the early Universe. These explosions likely marked the deaths of Population III stars and contributed to the chemical enrichment and dust content of the first galaxies. As facilities such as the Rubin Observatory and the Euclid, James Webb and Roman Space Telescopes begin to survey the dynamic and static high-redshift Universe (z ~ 2-10), identifying bona fide (P)PISNe becomes both crucial and increasingly challenging. Spectroscopic follow-up is limited, even for bright candidates, and several other explosion channels can mimic PISN-like signatures. This may lead to contaminated samples and biased event-rate estimates, with consequences for our broader understanding of stellar and galaxy evolution. In this talk, I will review insights from the most promising PPISN and PISN candidates of the past decade, emphasising the key observational challenges and theoretical uncertainties that affect their interpretation. I will discuss the multi-wavelength diagnostics and observing strategies needed to identify these rare transients in forthcoming surveys. I will also outline where current progenitor and explosion models fall short, and highlight areas where theoretical progress is urgently required to interpret candidates with confidence and to anticipate the types of PISN-like events that next-generation facilities may uncover. By combining these observational lessons with pressing theoretical questions, I will outline a path towards identifying genuine (P)PISNe and constraining the physics of the most extreme stellar explosions.

Modern astronomical facilities increasingly rely on tightly integrated hardware, software, and steady operations to deliver science-ready, high- quality data at scale. The Vera C. Rubin Observatory represents one of the most complex survey systems ever built, requiring comprehensive commissioning and system-level validation to transition from construction to science operations. In this talk, I will present key contributions to the commissioning and early operational phases of Rubin, with an emphasis on system performance analysis and the interface between instrumentation, observatory operations, and data quality. Drawing on several case studies—including investigations of camera sensor effects, mirror vibration, and dome thermal gradients—I will discuss how these investigations translated into measurable improvements in image quality, stability, and survey efficiency. Using blending and photometric redshift analyses as further examples, I will illustrate how observatory performance connects directly to downstream science measurements. I will conclude by discussing broader lessons from Rubin commissioning that are relevant to survey-driven facilities and collaborative instrumentation efforts.

Fast Radio Bursts (FRBs) are millisecond-duration radio transients from cosmological distances whose physical origins remain a mystery. With recent wide-field surveys, FRB research has entered a golden era in which large, homogeneous samples now allow statistically robust tests of long-standing questions: Are most FRBs intrinsically repeating? What controls burst emission properties? And how do progenitor environments shape the observed population? In this talk, I address these questions using theory-driven statistical approaches. First, using CHIME data, I show that the observed population of apparently non-repeating FRBs is consistent with weakly repeating sources, implying a high intrinsic repeater fraction and challenging the traditional repeater/non-repeater dichotomy. Second, focusing on repeating FRBs, I present time-frequency correlation analyses that reveal systematic spectral evolution in closely spaced burst pairs, providing new constraints on FRB emission mechanisms beyond single-burst analyses. Finally, I present the first systematic study of host-galaxy metallicities for localized FRBs, demonstrating how progenitor environments can be statistically constrained and linked to population-level properties. Together, these results illustrate how the golden era of FRB observations enables a unified understanding of FRB populations, emission physics, and host-galaxy environments, and opens new paths toward their application in cosmology.

Electronic liquid crystal (ELC) phases are spontaneous symmetry breaking states believed to arise from strong electron correlation in quantum materials such as cuprates and iron pnictides. Here, we report a direct observation of a smectic phase in a weakly correlated non-symmorphic Sb square-net semimetal, GdSbxTe2-x. Incommensurate smectic charge modulation and intense local unidirectional nanostructure, which coexist with Dirac fermions across Fermi level, are visualized by using spectroscopic imaging-scanning tunneling microscopy. As materials with highly mobile carriers are mostly weakly correlated, the discovery of such an ELC phase are anomalous and raise questions on the origin of their emergence. Specifically, we demonstrate how chemical substitution generates these symmetry breaking phases before the system undergoes a charge density wave (CDW)-orthorhombic structural transition [1]. Furthermore, the local unidirectional nanostructures appear coupled strongly with the CDW order. We will discuss the role of disorder in comparison with the recent claim of the bond density wave in CeSbTe [2]. Together, our results highlight the importance of impurities in realizing ELC phases and present a new material platform for exploring the interplay among quenched disorder, topology and electron correlation. 1. B. Venkatesan et al., npj Quantum Materials 10, 56 (2025). 2. X. Que et al., Nature Communications 16, 3053 (2025).

The Prime Focus Spectrograph and its Galaxy Evolution (PFS-GE) Survey mark a new era of spectroscopic programs, turning galaxy evolution into an information-rich, field-level frontier. We move from isolated measurements of individual objects and broad-stroke statistics to dense 3D maps of galaxies and their environments. The opportunity is enormous, but so is the inference challenge: the imprint of star formation, feedback, and gravitational evolution is nonlinear and entangled across scales. Much of this detail is washed away when we reduce the data to a small set of summary statistics like correlation functions. In this talk I’ll outline an end-to-end inference paradigm that combines differentiable forward modeling with field-level reconstruction of the cosmic density field, enabling complex datasets to be translated into physical constraints with principled uncertainty propagation. I’ll discuss how this approach naturally integrates multi-wavelength multi-prove observables and systematic effects within a single framework, and how it supports direct comparisons between competing physical models. Ultimately, this shifts us from simply fitting trends to reconstructing physical histories of the evolving interplay of structure, galaxies, and baryons.

The advent of JWST has revolutionized our understanding of the early Universe, most notably through the discovery of massive quiescent galaxies (QGs) at z>4. These systems challenge standard galaxy formation models, as they require both an extremely rapid assembly of stellar mass and an abrupt cessation of star formation within the first few billion years of cosmic time. To understand the mechanisms driving this early quenching, we must investigate the role of the environment. In this talk, I present a search for rare, high-redshift QGs within four proto-clusters in GOODS-S and GOODS-N fields at z~4-6 using JWST Cycle 1 data from the FRESCO program. Leveraging a comprehensive dataset—including up to 17 NIRCam filters, MIRI imaging, and ancillary HST data—we identify 13 QG candidates with stellar masses ranging from log(M_*/M_odot)~9-11. Our analysis reveals a clear positive correlation between stellar mass and the quiescent fraction. Notably, we observe significant variation across proto-clusters: proto-clusters in GOODS-N exhibit a 100% quiescent fraction in the highest mass bin log(M_*/M_odot)> 10.5), while analogous structures in GOODS-S appear devoid of massive QGs. Furthermore, these galaxies reside primarily on the outskirts of overdensity peaks—contrary to z~0 expectations—and we observe a sharp decline in QG prevalence at z > 5. These findings suggest that efficient quenching mechanisms were not yet fully mature at the highest redshifts and highlight the nuanced, diverse role of environment in early galaxy evolution.

Supermassive black holes are ubiquitous in the nearby Universe. Their lifecycle appears to be closely connected to the evolution of galaxies. How and when did these mysterious objects form? What was the first generation of black hole seeds? How did they grow quickly enough to power the most distant quasars? And how do black holes co-evolve with galaxies? I will summarize the demographics of central black holes in the local Universe and recent discoveries made with the James Webb Space Telescope that offer surprising, new insights into the earliest phases of black hole and galaxy formation during the first billion years of cosmic history. I will outline what we know and the much else that still remains uncertain.

Strong gravitational lensing is sensitive to the total mass distribution along the line of sight, making it a unique probe of dark matter in galaxies and clusters, and useful for studying resolved properties of the magnified background sources. Strong lensing is also valuable for constraining cosmology through lensed quasars and supernovae (SNe), which can be monitored to measure the "time delay" between the multiple images and constrain the Hubble constant (H_0). This method of measuring H0 is independent of the distance ladder and CMB observations, and may shed light on the H0 tension between early and late-Universe measurements. The latest results from the TDCOSMO project, including new JWST observations, constrains H_0 to be ~72 km/s/Mpc in a flat Lambda CDM cosmology with a precision of ~4.5%, while being maximally conservative with respect to galaxy mass profiles. As we move into the era of Rubin/LSST, new discoveries of lensed SNe will provide complementary constraints. The recently-discovered lensed SN "Winny" is the first galaxy-scale lensed SNe that could be useful for time-delay cosmography. I will present new high-resolution observations and preliminary lens models of this exciting system. Rubin/LSST will eventually discover ~thousands of lensed AGN and ~hundreds of lensed SNe, so finding these rare objects in such a huge dataset is a unique challenge. I discuss machine-learning search techniques that have been successfully applied to the HSC SSP in preparation for LSST and other large surveys such as Euclid and Roman. With these new observational facilities increasing the sample of lenses by over an order of magnitude, along with the resolving power and depth of JWST, we are poised to enter a new era of strong lensing cosmology.

Green valley (GV) galaxies represent systems transitioning from the star-forming to the quiescent phase and thus serve as ideal laboratories for probing the physical processes that suppress or shut down star formation. In this talk, I will first provide a broader perspective on why understanding galaxy quenching is important, and then present our efforts to address key questions using a combination of state-of-the-art integral field spectroscopy (IFS) surveys and radio observations enabled by the Atacama Large Millimeter/submillimeter Array (ALMA).

Exoplanet science encompasses one of the biggest existential questions: Are we alone in the universe? One promising approach to address this question is atmospheric spectroscopy, which provides a window into the composition and structure of exoplanet atmospheres. The recent launch of the James Webb Space Telescope and upcoming missions like Ariel and Twinkle provide a strong incentive for characterizing these distant worlds, with research into super-Earth and sub-Neptune exoplanets becoming increasingly popular because of their potential to reveal important clues into the origin and prevalence of life in the universe. However, the link between retrieved spectra and the underlying physical processes driving atmospheric enrichment is not well understood. Understanding these processes is crucial to interpret data and distinguish between biotic and abiotic spectral signatures. In this talk, I will explore the enrichment mechanisms leading to super-solar metallicities in super-Earth and sub-Neptune exoplanet atmospheres. The enrichment mechanisms are: (1) direct accretion of enriched nebular gas, (2) chemical processing between the planetary nucleus and the atmosphere, (3) preferential atmospheric escape, (4) geological outgassing, (5) late accretion of planetesimals, and (6) core erosion. My previous research has addressed the first three mechanisms, and my future work will explore the remaining three and identify population-level enrichment trends.

Supermassive black holes (SMBHs) are ubiquitous in the Universe, but their formation pathway remains one of the major open questions in astrophysics. One promising scenario for the origin of high-z SMBHs is the direct-collapse model. Recent JWST observations have also revealed a new population of sources, known as Little Red Dots (LRDs), at z = 4-8, which may be powered by accreting BHs. In this seminar, I will present results from our recent cosmological radiation-hydrodynamics simulations aimed to understand the origin of luminous quasars and LRDs in the early Universe. We performed zoom-in simulations of massive halos identified by Ishiyama and Hirano (2025) as a heavy BH seed formation site. Our simulations reveal the formation of overmassive BHs that grow to several times 10^7 Msun by z ~ 7, comparable to the BH masses inferred for high-z quasars. Remarkably, seed BHs with masses of order 10^6 Msun form about 10 kpc away from the main massive halo. After their formation, a brief super-Eddington accretion phase followed by Eddington-limited growth increases the BH mass beyond 10^7 Msun. We further performed higher-resolution zoom-in simulations around these BHs and found that a very dense gaseous disk forms around them, which may broaden the emission lines. We will discuss how this system resembles LRDs and what it implies for their origin.

The Jiangmen Underground Neutrino Observatory (JUNO) is a 20-kiloton liquid scintillator detector located 700 meters underground in Jiangmen, Guangdong, China. Its primary physics goal is to determine the neutrino mass ordering and to precisely measure the neutrino oscillation parameters by observing fine oscillation structures in the energy spectrum of reactor antineutrinos emitted from two nearby nuclear power plants at a baseline of 53 kilometers. After approximately twelve years of construction and commissioning, JUNO began physics data-taking in August 2025. Using an initial dataset corresponding to 59.1 days of high-quality data, JUNO has reported its first physics results, including precise measurements of two oscillation parameters: sin2

Primordial black holes (PBHs) are a possible dark-matter candidate that may have formed in the early Universe from the collapse of primordial density fluctuations. If a fraction of dark matter is in the form of PBHs, they can be searched for through gravitational microlensing. To test this possibility, we have been carrying out microlensing observations of stars in the Andromeda galaxy (M31) with the Subaru Hyper Suprime-Cam (HSC). In this talk, I will present the latest results from our Subaru HSC analysis based on a new pipeline designed to improve the sensitivity of the event search. The new pipeline adopts a more flexible microlensing light-curve model, including finite-source effects, which allows improved fits to candidate events. I will also discuss the interpretation of these candidates in light of recent independent reanalyses, and conclude with the implications for future PBH searches.

Exoplanetary atmospheres offer clues to their formation processes and atmospheric physics and chemistry under extreme conditions. In particular, recent JWST observations have revolutionized our understanding of atmospheric compositions. In the first half of my talk, I will introduce several key findings from JWST, which include multiple molecular detections such as SO2 and metal-rich atmospheres suggested for several sub-Neptunes, including the archetypal hazy sub-Neptune GJ1214b. I will also briefly highlight my recent work on cloud and haze microphysics. In the second half, I will discuss how we can interpret the JWST atmospheric observations based on planet formation models. In particular, I will examine how atmospheric mass–metallicity relations of giant exoplanets depend on their birthplaces and gas accretion processes. I will also highlight the potential need to revisit the interpretation of atmospheric sulfur, which has conventionally been attributed to planetesimal accretion.

The upcoming generation of wide and deep galaxy surveys, such as SPHEREx and Euclid, will provide a unique opportunity to probe the ultra-large scales of structure formation, where primordial non-Gaussianity, often parameterized by f_NL, is expected to leave its most detectable imprint. At the same time, these surveys inevitably enter regimes where relativistic and wide-angle effects become significant, requiring careful modeling to extract unbiased cosmological information. We address these challenges by applying the total-angular-momentum (TAM) formalism to describe redshift-space clustering beyond the flat-sky approximation. The standard Fourier-mode description, which characterizes distortions by the angle between a mode and a line of sight, becomes ill-defined over wide fields, whereas the TAM basis naturally separates radial and angular contributions and incorporates the relevant relativistic effects. Within this framework, we provide a new parameterization of wide-angle predictions that can be compared directly with survey data. We then introduce "PowerFull", a modification of the Julia-based "2-FAST" package, which enables efficient computation of the full relativistic angular power spectrum. Finally, using Fisher information matrix–based analyses, we show that neglecting these effects biases parameter estimation, and that achieving O(1) precision on f_NL with surveys like SPHEREx requires a consistent relativistic and wide-angle treatment of galaxy clustering.

Are we alone? You often hear this question because the formation of a terrestrial planet had been, theoretically, considered difficult if not impossible. This was changed after the NASA Kepler and K2 missions found that rocky planets are ubiquitous. Afterwards, it has become popular to resolve the natal environment of rocky planets, the protoplanetary disks, and then investigate the planet-formation activities observationally/theoretically. Without a surprise, most of those studies concluded that the protoplanetary disks we resolved are efficiently forming rocky planets, which, however, yield planets that are too wet and too carbonaceous to compare with our own Earth. Are we not alone? This is the issue I have been tackling. In this talk, I will provide an overview of this research field and outline our contribution/destruction.

This paper critically examines the concept of transparency in AI systems within the domains of law and governance using tools in philosophy. Current discussions often treat transparency, explainability, and trust as closely linked, frequently requiring explainable AI (XAI) techniques as necessary justifications of AI decisions. Drawing on philosophy, cognitive science, and jurisprudence, we argue in this paper that this approach involves a category mistake: XAI explanations typically describe causal mechanisms or computational processes, which are levels below that of propositional justification expected in human deliberation, including human legal and theoretical reasoning. It concludes that transparency should be understood as a requirement in social dilbeberation rather than a purely technical feature of AI systems. Accordingly, the paper proposes a pluralistic, context-sensitive approach to AI transparency informed by analogies with existing legal and regulatory practices, with XAI serving as a useful but neither necessary nor sufficient tool.

Stellar ages are notoriously difficult to measure accurately for main-sequence low-mass stars. This limits our ability to address questions ranging from the evolutionary state of exoplanets to the chemical history of the Galaxy. Gyrochronology, which uses stellar rotation as a proxy for age, is a promising solution to this quandary. Unfortunately, however, we are still far from being able to describe fully the evolution of rotation for low-mass stars, or from being able to use rotation measurements to estimate accurately the ages of isolated field stars. I will summarize recent ground-based and space-based work to characterize the rotational behavior of G, K, and M dwarfs in open clusters ranging in age from 125 Myr (the Pleiades) to 3 Gyr (Ruprecht 147), and then compare these data to each other and to models for stellar spin-down to appraise our current understanding of the age-rotation relation. I will also touch on an on-going survey targeting newly discovered single-aged stellar populations to address the broader age-rotation-activity relation.

The advancement of astronomical instrumentation in the last century has enhanced our understanding of the universe by enabling more precise observations of new astrophysical objects. Yet, there remains a growing need for larger apertures to achieve greater light-collecting ability, and higher-resolution spectroscopic measurements to resolve finer details of stars and planets. The concept of combining light from individual telescopes, by using fused silica fibers to create a larger effective aperture, was proposed by Angel et al. in 1977. Building on this idea, the Large Fiber Array Spectroscopic Telescope (LFAST) project seeks to construct large arrays of small, individual fiber-fed telescopes to feed high-resolution spectrographs. Our goal is to replicate a system consisting of 20 unit telescopes, forming an array with a collecting area comparable to that of the planned Extremely Large Telescope, at a fraction of its cost.

Precursors of core-collapse supernovae are direct probes of the violent mass-loss episodes that massive stars undergo shortly before core collapse. While long-lasting (~10 to >70 days) precursors of hydrogen-poor supernovae (such as Type Ibn) have been observed, stellar eruption models struggle to explain these extended durations given the compact nature of their stripped-envelope progenitors. In this talk, I will propose a "shock-powered transient" scenario to resolve this discrepancy, in which the kinetic energy of erupted material is dissipated through collision with pre-existing circumstellar matter. Our multi-band light-curve model successfully reproduces the luminosities and durations of observed precursors (SNe 2006jc, 2019uo, 2021foa, and 2022pda). Combined with hydrodynamical mass-eruption simulations using MESA and CHIPS, we constrain the progenitors to be compact, hydrogen-poor massive stars with M_ZAMS ≳ 25 M_sun. Finally, I will discuss the eruption mechanisms and the core-collapse explosion process of these supernovae.

In the last few years, photomety and spectroscopy from ground and space have opened up an exciting domain: accreting super-Jupiters, in several cases found in a gap around their parent star. This raises hopes of sharpening our understanding of processes shaping planet formation. The accretion shocks at the surface of a planet and its circumplanetary disc can be sources of hydrogen-line emission, detected at several planetary-mass accretors. Separately, if a young forming planet has a strong magnetic field, this field might channel the gas from the vicinity of the planet onto it, as for young stars. This too should lead to line emission. We present an overview of recent (non-)detections in this rapidly-moving field and review the relevant theory. (1) We discuss how JWST can study even faint accretion tracers at planetary-mass objects. We put this in context of the not mutually exclusive accretion frameworks. (2) We present a semi-analytical model yielding line emission from the multidimensional accretion onto a planet. It reveals that accreting planets are much fainter than expected from simple formulae, which could be a major factor explaining the scarcity of detections so far. (3) We also present predictions of high-resolution profiles of hydrogen lines and their observability for different objects. We focus on Brackett alpha, accessible to the upcoming ELT/METIS (R=100,000) spectrograph.

Supernova remnants (SNRs) have been considered as a source of the bulk of Galactic cosmic rays (CR). Recently, some evidence of extended gamma-ray emission have been reported by e.g. LHAASO and H.E.S.S. and this emission often has a high maximum energy, suggesting the existence of particles escaping from acceleration regions of SNRs. Although some analytical works have explained based on the escaped CR, and have calculated the gamma-ray spectrum on the environment with SNR shock interacting with clumps and have shown the spectral hardening, such works have generally not included the CR propagation. Therefore, we revisit a setup similar to the conventional CR propagation calculation, which solved the time evolution of the spatial distribution and spectrum of the accelerated particles of evolving shock and the behavior of the evolution. In this work, we developed a simulation code that calculates the time evolution of phase-space distribution function of CR spectrum by solving the one-dimensional diffusion-convection equation coupled with analytical solutions for shock dynamics. Our results show that the total energy of accelerated particles still increases even after the age of 10kyrs in our models. We also calculated Pi0 gamma-ray emission at various stages of SNR evolution with/without clumps located at shock upstream, and compare them with several observed objects. The results show that the emission from clumps in the vicinity can dominate the total flux, depending on the distance between the shock and the clumps. To explain the observation data, the existence of clumps can play an important role.

A decade has passed since the first ALMA detection of [O III] 88 μm emission at the epoch of reionization (EoR), which has greatly advanced our understanding of galaxy formation up to z = 14. Also, the [CII] 158 μm line has been established as a powerful probe of neutral gas contents and star formation activities in EoR galaxies. Yet, such far-infrared nebular line studies remain largely limited to the southern sky due to ALMA's location. We present LMT-FINEST, a new program to explore far-infrared nebular emission using FINER, an SIS mixer receiver to be deployed on the Large Millimeter Telescope (LMT) in Mexico in late 2026. Building on ALMA's Wideband Sensitivity Upgrade (WSU) technology, FINER will cover ALMA Bands 4-7 with an instantaneous frequency span of > 30 GHz per polarization, delivering Δz ~ 1 at z ~ 10 in a single tuning. At 40% of ALMA’s light-collecting area, the LMT's similar atmospheric conditions and FINER's 5 times wider bandwidth compared to ALMA offer an unparalleled spectral scanning capability in the northern hemisphere, paving the way for efficient detection of [O III]/[C II] in galaxies identified by JWST and Euclid, and soon by Roman.

Close-in, small exoplanets around M-dwarf stars have revealed a remarkable variety of worlds, ranging from rocky super-Earths to volatile-rich sub-Neptunes. However, understanding how these planets acquire their volatile ingredients and hold onto their atmospheres in tightly packed configurations remains a major puzzle. In compact systems, tight orbital spacing and multi-body interactions drastically alter how planets accrete volatiles, leading to atmospheric outcomes that are distinct from those in wider-orbit systems such as our own.
This talk presents preliminary efforts in understanding the atmospheric and volatile lifecycle of exoplanets in these extreme environments. The discussion will walk through three major stages of planetary evolution: i) The initial delivery and accumulation of volatiles while the planet is still forming, ii) the thermochemical evolution and mass loss from primordial gaseous envelopes, and iii) the long-term climate dynamics and chemistry that drive the emergence of secondary atmospheres and their spectral signatures.

The star- and disk-formation process involves balancing gravity, turbulence, and magnetic fields. Turbulence and magnetic fields attempt to slow down the gravitational collapse. Thanks to high-angular resolution interferometric observations, much progress has been made in studying the disk properties. Unfortunately, little progress has been made in understanding the role of the environment in the formation and evolution of stars and disks. In my talk, I will present the main results from ProPStar and PRODIGE, focusing on the connection between the physical and chemical properties of star-forming environments and the evolution of protoplanetary disks. By combining observations across different environments with modelling, these projects provide new constraints on how initial conditions shape disk structure, composition, and the material available for planet formation. I will discuss what these results reveal about the link between environment and disk, and outline prospects for future observational and modelling efforts aimed at tracing how the earliest stages of star formation influence the emergence of planetary systems.

The Circinus galaxy hosts one of the nearest active galactic nuclei (AGN), offering a unique opportunity to study the structure of the AGN torus, black hole accretion, and feedback processes in unprecedented detail. In this talk, I present a comprehensive multiwavelength investigation of the Circinus nucleus, focusing on the physical connection between cold and warm gas, accretion, and outflows. First, using high-angular-resolution ALMA observations, we resolve the spatial distribution and kinematics of multiphase gas in the central region. We find that molecular, atomic, and ionized gas components exhibit distinct dynamical behaviors, collectively forming a complex, multiphase AGN torus rather than a single coherent structure. Furthermore, absorption-line observations enable us to detect dense inflowing gas toward the AGN. The derived physical conditions indicate that gravitational instability of the parsec-scale dense disk plays a key role in driving gas accretion from parsec scales down to the vicinity of the supermassive black hole. Second, we present an analysis of JWST MIRI/MRS data to investigate the properties of the warm AGN torus. In addition to directly estimating the molecular gas mass by using multiple H2 lines, we detect high-ionization emission lines (e.g., coronal lines), which reveal fast outflows on parsec scales. By combining these results with ALMA observations of the cold torus, we discuss the AGN torus from a truly multi-phase and multi-scale perspective, highlighting the interplay between accretion and feedback in a nearby benchmark AGN.

Galactic Circumnuclear regions, where billions of stars and putative supermassive black holes (SMBHs) interact intensely with the surrounding interstellar medium (ISM), plays a crucial role in our understanding of the feeding and feedback of SMBHs, as well as the co-evolution of the central black hole and the host galaxy. However, the direct connection between black hole feedback and the properties of the surrounding ISM remains elusive. This work makes use of multiwavelength observations to study the multiphase ISM in the circumnuclear regions of nearby galaxies (M31 and M81) in order to understand the details of this interaction. These closest galaxies provide us with a full-band, spatially-resolved view of galactic nuclei, serving as unique laboratories for understanding SMBH feedback. Through millimeter and infrared observations, it has been found that the cold gas is severely deficient in the circumnuclear region of M31 compared to that in the Milky Way. Nevertheless, the cold gas temperature is much higher than that of the galactic disk, which may be related to past nuclear activities. In addition, it is found that the nuclear region exhibits a much higher Virial parameter than the galactic disk, suggesting that the gas clump in this region is rather pressure bound or transient. The notable distinction in the molecular gas environment between the nuclear region and the disk carries significant implications (e.g., a minor mereger) for the physical conditions within this particular region. On the other hand, through optical integrated field spectroscopic observations, it is also found that there is a biconical ionized gas outflow structure located at about 200 pc from M81 center, which may be directly related to the low luminosity active galactic nucleus in M81. We use CLOUDY photoionization model to quantitatively describe the LINER radiation in the central region of both M31 and M81, and found a deficiency of [OIII] lines, indicating additional ionization sources may play a role.

Interpretable machine learning (ML) techniques and artificial intelligence (AI) are revolutionizing our ability to study galaxy evolution and large-scale structure. Convolutional neural networks (CNNs) can now reliably predict galaxies' physical properties, including cold gas content and metallicity, directly from three-color optical images. These models can even reconstruct entire optical spectra from imaging alone. Highly optimized CNNs can also robustly identify nearby dwarf galaxies from wide-area surveys, expanding the sample of known low-redshift satellite systems by over 10-fold. Meanwhile, graph neural networks (GNNs) can encode simulated galaxies amid their surroundings, learning how the galaxy–halo connection varies with large-scale environment. These applications demonstrate how explainable ML models with strong inductive biases enable new scientific insights in galaxy evolution and cosmology. In the era of wide-area galaxy surveys by the Vera C. Rubin Observatory, Nancy Grace Roman Space Telescope, and Euclid, advanced ML and interpretable AI methods will play an increasingly prominent role in extracting physical understanding from astronomical datasets.

Roughly half of all starlight ever emitted has been absorbed by interstellar dust and re-radiated at far-infrared wavelengths, fundamentally shaping our view of galaxy evolution. Even at 600 million years after the Big Bang, galaxies harbour warm dust reservoirs rivaling the brightest local systems. Yet despite the high stellar masses revealed by JWST (implying millions of dust-producing supernovae), deep ALMA observations detect no dust emission, highlighting our uncertainties in the emergence, evolution and impact of dust across the past 13.8 billion years. I will present an overview of our best constraints on early obscured galaxy evolution. Using ALMA Bands 1 through 10 efficiently, I detail our best picture on resolved dust, gas and star formation maps in 200 dusty star-forming galaxies, revealing the diversity of extreme star formation modes. Furthermore, I systematically investigate their larger environments through a novel machine learning framework, exploring the dust-selected protocluster candidates from substantial archival datasets (>2000 hours) combined with ongoing spectroscopic programs. Extending dust studies to younger galaxies, I will present a picture of dust formation and emission properties from composite ALMA imaging out to the pre-reionization Universe, concluding with recommendations for next-generation facilities (AtLAST, ALMA2040).

This talk presents an analytic description for the late stages of giant planet formation, when planets gather the majority of their mass. The resulting solutions show how the protoplanet properties (envelope density distribution, velocity field, column density, disk surface density, system luminosity, and emergent spectral energy distributions) vary with the input parameters of the problem (instantaneous mass, orbital location, accretion rate, and planetary magnetic field strength). We then construct a framework for calculating the distribution of planet masses resulting from this paradigm. In this scenario, the disk lifetime determines the end of mass accretion onto the planet. The mass accretion rate depends on the size of the Hill sphere, the fraction of the disk accretion flow that enters the sphere of influence, and the efficiency with which the planet captures the incoming material. The resulting model produces a planetary mass function with a nearly power-law form, roughly consistent with current observational estimates.

Accretion onto compact objects is one of the fundamental mechanisms for powering high energy emission in quasars, active galactic nuclei, and X-ray binaries. In this talk, I will present our study of relativistic accretion flow around various compact objects and discuss its astrophysical significance. We investigate relativistic accretion flows in the strong gravity regime around compact objects by incorporating the effective potential. The infalling rotating matter may encounter a centrifugal barrier, and depending on the shock conditions, a post-shock corona (PSC) of hot, dense electrons can form. PSC reprocesses the soft disk photons to high energy radiations via inverse Comptonization. We show that shock-induced accretion flows around Kerr-Taub-NUT BHs are thermodynamically favored and yield significantly higher luminosity than shock-free solutions. BHs are not the only possible accreting compact objects, exotic compact objects like wormholes (WHs) remain theoretically viable. Therefore, we investigate the relativistic accretion flow around Kerr-like WH and show that the accretion solutions around WH can explain the observed luminosity of Cygnus X-3 in its hypersoft state. In this framework, we further examine the impact of dark matter (DM) halos on accretion dynamics, considering that the BH is surrounded by a Hernquist, Navarro-Frenk-White (NFW), Einasto, or DM spike profile. DM shifts the critical points inward and enhances disk luminosity, which offers a potential probe of DM distribution near BHs. Finally, within a steady-state GRMHD framework, we model magnetized accretion around Kerr BHs motivated by Event Horizon Telescope observations of Sagittarius A*. Our solutions reproduce magnetic field strengths within ±10% of the EHT-constrained values, that provides a consistent theoretical description of magnetized accretion flows.

Planets form in disks of gas and dust around young stars, but when and how they emerge remains uncertain. ALMA observations have revealed that protoplanetary disks are rich in rings, gaps, spirals, and other substructures, especially in dust continuum emission, yet interpreting their physical origins is challenging. In this talk, I will discuss how observations of disk gas dynamics can provide a complementary approach to probing the physical conditions in disks. I will focus on exoALMA, an ALMA Large Program designed to map the three-dimensional kinematic structure of 15 nearby protoplanetary disks at high spatial and spectral resolution. The survey reveals widespread deviations from smooth Keplerian disk, including large-scale spiral-like patterns and vertical gas motions. I will present an overview of these diverse kinematic structures, which will serve as a useful basis for future, more detailed investigations of disk dynamics and planet formation processes.

Fast Radio Bursts (FRBs) are bright, millisecond pulses of radio emission of extragalactic origin, but their physical origins are still not well-understood. The Canadian Hydrogen Intensity Mapping Experiment FRB (CHIME/FRB) project has ushered in a new era of FRB science, now delivering subarcsecond-precision localizations through the addition of Outrigger stations, and their growing sample of precisely localized FRBs enables characterization of their robustly-associated host galaxies. I analyze the full sample of FRB hosts identified from both the first and second CHIME/FRB Outrigger catalogs, constructing a uniform dataset and enabling robust constraints on galaxy properties across a wide range of redshifts and environments. I place this expanded and uniformly analyzed population of hosts in the broader context of the known FRB host galaxy distribution and other transient host distributions, providing new insight into the diversity of FRB environments and their implications for progenitor models. Finally, I will discuss a second ongoing project investigating possible connections between the burst rates of repeating FRBs and their host properties.

The Clusters Hiding in Plain Sight (CHiPS) survey uncovered CHIPS1911+4455 at z = 0.485, a unique system that combines one of the lowest-entropy, fastest-cooling cores known with a highly disturbed X-ray morphology. At its center, the brightest cluster galaxy is undergoing a massive starburst (~150 Msun yr⁻¹), making it one of the most extreme BCGs at low redshift and a rare analog of high-z star-forming systems. I will present the discovery and characterization of this cluster with Chandra and HST, and briefly highlight recent multiwavelength follow-up, including NOEMA detections of a large molecular gas reservoir and JVLA/VLBA evidence for a newly born radio AGN. Together, these results suggest that CHIPS1911+4455 is caught in a rare transitional phase, where merger-driven turbulence and runaway cooling fuel both intense star formation and the onset of AGN feedback.

Carbon abundances in rocky planets play an important role in shaping their surface environments. Additionally, the C/O ratio in the atmospheres of gas giant exoplanets may reveal their birthplace. Therefore, it is crucial to investigate how carbon is distributed in both the gas and solid phases of protoplanetary disks. However, despite active discussion in recent years, this remains unclear due to uncertainties about the forms of carbon-bearing species in the solid phase and the motion of solids in protoplanetary disks. In this talk, I will present the current understanding of carbon abundance and its time evolution in protoplanetary disks for both the gas and solid phases. In the first half, I will address the unresolved problem of the solid phase: the Earth’s carbon deficit. The Earth is significantly depleted in carbon compared to comets and interstellar dust, which contain abundant refractory carbonaceous solids, such as organics and amorphous hydrocarbons. Since the disk temperature around Earth’s orbit is not high enough to vaporize these carbonaceous solids, this significant depletion remains an unsolved problem. I will discuss several models proposed to explain this deficit, focusing on thermal decomposition of organics and gasification by far-ultraviolet radiation (photolysis), as well as the limitations of these models. In the second half, I will examine the carbon abundance in the gas phase of disks. While classical disk models only considered the sublimation of volatile carbon species such as CO and CO₂, recent studies have shown that degassing from refractory carbonaceous solids significantly contributes to the gas-phase carbon abundance. I will focus on the effects of degassing caused by photolysis and thermal decomposition of these solids.

Intensity interferometry exploits temporal correlations in photon arrival times arising from the statistical properties of light to achieve angular resolution far exceeding conventional methods. Recent advances in photon-counting detectors and signal processing are enabling a new generation of intensity interferometers capable of high-precision astronomical measurements. I will present two techniques that broaden the scientific reach of this approach. First, the expanding ejecta method (EEM) uses intensity interferometric measurements of supernovae to determine angular diameter distances from purely geometric arguments, providing a new rung on the cosmic distance ladder or enabling direct inference of the Hubble constant independent of standard candles. Second, extended-path intensity correlation (EPIC), a variant of intensity interferometry, enables differential astrometry at micro-arcsecond precision over fields of view spanning several arcseconds. One flagship can achieve differential astrometry at micro-arcsecond precision on sources with high surface brightness over fields of view spanning several arcseconds. One flagship application is the measurement of astrometric lensing perturbations in multiply imaged quasars caused by extremely small dark matter structures. Such measurements could probe sub-stellar-mass halos, thereby providing a sensitive probe of dark matter microphysics inaccessible to other observational methods.

As physicists begin to embed LLMs and agents into analysis workflows, the promise of an automated discovery pipeline is growing. Yet these pipelines still rely on "reconstruction"—the extraction of high-level physics features from raw sensor data (e.g., waveforms, images) using domain-specific tools. Because reconstruction targets a predefined set of features, it can act as an information bottleneck and introduce bias when the underlying physics model is imperfect; for simulation-trained supervised models, this appears as data shift. We address these issues with a domain-specific Foundation Model (DSFM) that learns physics representations directly from raw, sensor-level data via self-supervised learning. The resulting representations are reusable across downstream reconstruction and analysis tasks and are free from simulation mis-modeling. I will present our sensor-level DSFM across multiple detector technologies in High Energy Physics, showing reconstruction performance competitive with or better than state-of-the-art models and up to 10,000× greater label efficiency. Finally, I show that fine-tuning can also be driven directly by data using a differentiable physics simulator.

Lyman-break galaxies (LBGs) are star-forming galaxies selected via the photometric "dropout" caused by the Lyman-break feature redshifting through optical passbands at z > 2, making them an efficient tracer of the high-redshift universe. The Subaru Strategic Program ʻŌnohiʻula Prime Focus Spectrograph (PFS) Galaxy Evolution Survey is about to begin its science operations, with its Tomographic sub-working group targeting LBGs at z = 2.3–3.4 to reconstruct the matter density field via the Lyman-alpha forest. In this talk, I will present a parallel effort to constrain the galaxy bias and growth-rate of large-scale structure directly from the galaxy-galaxy clustering of these same LBGs. I will focus on the construction of a realistic LBG lightcone mock, built to validate our clustering measurement and modeling pipeline and to characterize PFS-specific observational systematics. This mock-driven approach will let us deliver a robust, state-of-the-art constraint on cosmic growth-rate, fσ₈, at the highest redshifts currently probed by galaxy clustering.

We propose a next-generation international ground-based gravitational-wave detector in Taiwan, the Cryogenic sub-Hz cROss torsion-bar detector with quantum NOn-demolition Speed meter (CHRONOS), optimized for the unexplored 0.1–10 Hz frequency band between the space-based detector LISA and future ground-based detectors such as Cosmic Explorer and the Einstein Telescope. CHRONOS combines a ring-cavity Sagnac interferometer with torsion-bar test masses to realize the first quantum nondemolition (QND) measurement of angular momentum in a macroscopic system. By implementing a speed-meter readout in the rotational degree of freedom, CHRONOS coherently suppresses quantum radiation-pressure noise, enabling observations in the sub-Hz regime. We have also proposed a novel detuned power-recycling scheme to further optimize the detector sensitivity. A distinctive feature of CHRONOS is its compact design. Unlike kilometer-scale gravitational-wave observatories, the experiment can be implemented within a footprint of approximately 10 m × 10 m. This compact configuration makes it possible to develop and operate a sub-Hz gravitational-wave detector in a conventional underground laboratory while retaining sensitivity to a wide range of astrophysical, cosmological, and geophysical signals. Based on a realistic optical design incorporating torsion-bar test masses, we estimate a strain sensitivity of approximately (h ~ 1 x 10^{-18} 1/√Hz) at 2 Hz with 2.5 m arm lengths. This sensitivity enables (i) the direct detection of intermediate-mass black hole binaries, (ii) searches for stochastic gravitational-wave backgrounds, (iii) tests of Yukawa-type deviations from gravity, and (iv) searches for dark matter. Furthermore, even the 2.5 m prototype is expected to enable the prompt detection of gravity-gradient signals from earthquakes with magnitudes of approximately M5.5. CHRONOS therefore opens new opportunities for quantum-limited geophysical observations as well as multi-band and multi-messenger gravitational-wave astronomy. In this talk, we will present the current status of the CHRONOS project and recent progress in its international collaboration.

The Giant Magellan Telescope, of which ASIAA is a Founder member, is one of two new-generation extremely large telescopes currently under construction. It is a public-private partnership between a consortium of 16 leading research institutions and the US National Science Foundation. The GMT, with its innovative design and capable instrumentation, will help us to answer key questions about the evolution of galaxies in the early universe, the growth of supermassive black holes and the frequency of habitable planets. I will discuss the key science issues and how the GMT’s capabilities address them, describe the status of the project and talk about plans for the near-term future and about recent accomplishments.

Cosmological large-scale structures are shaped by nonlinear processes mainly driven by gravity. Previous analysis methods rely on theoretical templates that are based on perturbative expansion about the linear solution, restricting the extraction of information to large, mildly nonlinear, scales. On the other hand, N-body simulations can uncover structures on smaller scales, until non-gravitational effects such as gas cooling and feedback processes eventually become a factor. However, their high computational cost hinders the direct use in statistical inference. In this talk, I will discuss the emulator approach as a potential solution. In particular, I will present our Dark Quest simulation project, especially the updated simulation code GINKAKU and the matter power spectrum emulator from the second round of the campaign. The core idea lies in the flexibility of neural network models. We show how the additional inputs to the emulator, such as the initial phases, resolution parameters, and linear power spectra, improve the predictions. I conclude by comparisons with previous emulators from independent groups, highlighting the current state of the emulator approach.

The Sgr A* black hole is along a hyper strong scattering sightline. I present time resolved VLBA/HSA of the nearby magnetar, indicating two scattering screens: one near the earth, one half way. This places new constraints on the magnetic environment of Sgr A*, velocities in the ISM, and scattering structures. The polarization requires a third screen close to the black hole. A novel component of this analysis for me is the extensive use of AI agents, which accelerated the analysis hundredfold. I describe the experience, and opportunities for the future.

The James Webb Space Telescope has discovered a population of so-called "Little Red Dots" at high redshift. These are very compact astronomical sources with a red colour and a V-shaped spectrum, with the minimum located at the Balmer break. In this talk, I will review the possible interpretations of these sources, including dusty starburst galaxies, Active Galactic Nuclei and possible alternative configurations, including gas-enshrouded black holes. I will discuss the implications of the different possibilities and show how many of them imply that an active black hole is already there or in the process of formation. Subsequently, we will discuss the environments in which such sources may form. For this purpose, I will present the results of dedicated high-resolution cosmological simulations in the regime of high star formation efficiencies, a property that very likely is required to explain the shallowness of the galaxy mass function obtained with JWST. I will show how such environments very naturally produce dense and compact configurations including short migration timescales for the gas and star clusters, thereby producing a natural pathway to form a central massive black hole.