Search

Thanks to the rapidly increasing time-domain facilities, we are entering a golden era of research on gamma-ray bursts (GRBs). In this Letter, we report our observations of GRB 240529A with the Burst Optical Observer and Transient Exploring System, the 1.5-meter telescope at Observatorio Sierra Nevada, the 2.5-meter Wide Field Survey Telescope of China, the Large Binocular Telescope, and the Telescopio Nazionale Galileo. The prompt emission of GRB 240529A shows two comparable energetic episodes separated by a quiescence time of roughly 400 s. Combining all available data on the GRB Coordinates Network, we reveal the simultaneous apparent X-ray plateau and optical re-brightening around $10^3-10^4$ s after the burst. Rather than the energy injection from the magnetar as widely invoked for similar GRBs, the multi-wavelength emissions could be better explained as two shocks launched from the central engine separately. The optical peak time and our numerical modeling suggest that the initial bulk Lorentz factor of the later shock is roughly 50, which indicates that the later jet should be accretion-driven and have a higher mass loading than a typical one. The quiescence time between the two prompt emission episodes may be caused by the transition between different accretion states of a central magnetar or black hole, or the fall-back accretion process. A sample of similar bursts with multiple emission episodes in the prompt phase and sufficient follow-up could help to probe the underlying physics of GRB central engines., Comment: Resubmitted to ApJL after addressing the referee's comments; comments are welcome

In this study, we investigate whether non-English-centric LLMs, despite their strong performance, `think' in their respective dominant language: more precisely, `think' refers to how the representations of intermediate layers, when un-embedded into the vocabulary space, exhibit higher probabilities for certain dominant languages during generation. We term such languages as internal $\textbf{latent languages}$. We examine the latent language of three typical categories of models for Japanese processing: Llama2, an English-centric model; Swallow, an English-centric model with continued pre-training in Japanese; and LLM-jp, a model pre-trained on balanced English and Japanese corpora. Our empirical findings reveal that, unlike Llama2 which relies exclusively on English as the internal latent language, Japanese-specific Swallow and LLM-jp employ both Japanese and English, exhibiting dual internal latent languages. For any given target language, the model preferentially activates the latent language most closely related to it. In addition, we explore how intermediate layers respond to questions involving cultural conflicts between latent internal and target output languages. We further explore how the language identity shifts across layers while keeping consistent semantic meaning reflected in the intermediate layer representations. This study deepens the understanding of non-English-centric large language models, highlighting the intricate dynamics of language representation within their intermediate layers., Comment: work in progress

Hot dust-obscured galaxies (Hot DOGs) are a population of hyperluminous, heavily obscured quasars discovered by the \emph{Wide-field Infrared Survey Explorer} (\emph{WISE}) all-sky survey at high redshift. Observations suggested the growth of these galaxies may be driven by mergers. Previous environmental studies have statistically shown Hot DOGs may reside in dense regions. Here we use the Very Large Telescope (VLT) narrowband and broadband imaging to search for Ly$\alpha$ emitters (LAEs) in the 6.8' * 6.8' field of the Hot DOG W2246$-$0526 at $z=4.6$. W2246$-$0526 is the most distant Hot DOG. We find that there is an overdensity of LAEs in W2246$-$0526 field compared with the blank fields. This is the direct evidence that this most distant Hot DOG is in an overdense environment on the Mpc scale, and the result relates to the merger origin of Hot DOGs., Comment: 11 pages, 8 figures, Accepted for publication in The Astrophysical Journal

Alzheimer's Disease (AD) is an irreversible neurodegenerative disorder that often progresses from Mild Cognitive Impairment (MCI), leading to memory loss and significantly impacting patients' lives. Clinical trials indicate that early targeted interventions for MCI patients can potentially slow or halt the development and progression of AD. Previous research has shown that accurate medical classification requires the inclusion of extensive multimodal data, such as assessment scales and various neuroimaging techniques like Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET). However, consistently tracking the diagnosis of the same individual over time and simultaneously collecting multimodal data poses significant challenges. To address this issue, we introduce GFE-Mamba, a classifier based on Generative Feature Extraction (GFE). This classifier effectively integrates data from assessment scales, MRI, and PET, enabling deeper multimodal fusion. It efficiently extracts both long and short sequence information and incorporates additional information beyond the pixel space. This approach not only improves classification accuracy but also enhances the interpretability and stability of the model. We constructed datasets of over 3000 samples based on the Alzheimer's Disease Neuroimaging Initiative (ADNI) for a two-step training process. Our experimental results demonstrate that the GFE-Mamba model is effective in predicting the conversion from MCI to AD and outperforms several state-of-the-art methods. Our source code and ADNI dataset processing code are available at https://github.com/Tinysqua/GFE-Mamba., Comment: 35 pages, 4 figures

This paper introduces LLM-jp, a cross-organizational project for the research and development of Japanese large language models (LLMs). LLM-jp aims to develop open-source and strong Japanese LLMs, and as of this writing, more than 1,500 participants from academia and industry are working together for this purpose. This paper presents the background of the establishment of LLM-jp, summaries of its activities, and technical reports on the LLMs developed by LLM-jp. For the latest activities, visit https://llm-jp.nii.ac.jp/en/.

In terms of the variable nature of normal active galaxy nuclei (AGN) and luminous quasars, a so-called flux variation gradient (FVG) method has been widely utilized to estimate the underlying non-variable host galaxy fluxes. The FVG method assumes an invariable AGN color, but this assumption has been questioned by the intrinsic color variation of quasars and local Seyfert galaxies. Here, using an up-to-date thermal fluctuation model to simulate multi-wavelength AGN variability, we theoretically demonstrate that the FVG method generally overestimates the host galaxy flux; that is, it is more significant for brighter AGN/quasars. Furthermore, we observationally confirm that the FVG method indeed overestimates the host galaxy flux by comparing it to that estimated through other independent methods. We thus caution that applying the FVG method should be performed carefully in the era of time-domain astronomy.

In our increasingly interconnected world, where intelligent devices continually amass copious personalized multi-modal data, a pressing need arises to deliver high-quality, personalized device-aware services. However, this endeavor presents a multifaceted challenge to prevailing artificial intelligence (AI) systems primarily rooted in the cloud. As these systems grapple with shifting data distributions between the cloud and devices, the traditional approach of fine-tuning-based adaptation (FTA) exists the following issues: the costly and time-consuming data annotation required by FTA and the looming risk of model overfitting. To surmount these challenges, we introduce a Universal On-Device Multi-modal Model Adaptation Framework, revolutionizing on-device model adaptation by striking a balance between efficiency and effectiveness. The framework features the Fast Domain Adaptor (FDA) hosted in the cloud, providing tailored parameters for the Lightweight Multi-modal Model on devices. To enhance adaptability across multi-modal tasks, the AnchorFrame Distribution Reasoner (ADR) minimizes communication costs. Our contributions, encapsulated in the Cloud-Device Collaboration Multi-modal Parameter Generation (CDC-MMPG) framework, represent a pioneering solution for on-Device Multi-modal Model Adaptation (DMMA). Extensive experiments validate the efficiency and effectiveness of our method, particularly in video question answering and retrieval tasks, driving forward the integration of intelligent devices into our daily lives.

Deep learning models have performed well on many NLP tasks. However, their internal mechanisms are typically difficult for humans to understand. The development of methods to explain models has become a key issue in the reliability of deep learning models in many important applications. Various saliency explanation methods, which give each feature of input a score proportional to the contribution of output, have been proposed to determine the part of the input which a model values most. Despite a considerable body of work on the evaluation of saliency methods, whether the results of various evaluation metrics agree with human cognition remains an open question. In this study, we propose a new human-based method to evaluate saliency methods in NLP by crowdsourcing. We recruited 800 crowd workers and empirically evaluated seven saliency methods on two datasets with the proposed method. We analyzed the performance of saliency methods, compared our results with existing automated evaluation methods, and identified notable differences between NLP and computer vision (CV) fields when using saliency methods. The instance-level data of our crowdsourced experiments and the code to reproduce the explanations are available at https://github.com/xtlu/lreccoling_evaluation., Comment: 13 pages, 4 figures, Accepted for LREC-Coling 2024 (Oral)

The creation of instruction data and evaluation benchmarks for serving Large language models often involves enormous human annotation. This issue becomes particularly pronounced when rapidly developing such resources for a non-English language like Japanese. Instead of following the popular practice of directly translating existing English resources into Japanese (e.g., Japanese-Alpaca), we propose an efficient self-instruct method based on GPT-4. We first translate a small amount of English instructions into Japanese and post-edit them to obtain native-level quality. GPT-4 then utilizes them as demonstrations to automatically generate Japanese instruction data. We also construct an evaluation benchmark containing 80 questions across 8 categories, using GPT-4 to automatically assess the response quality of LLMs without human references. The empirical results suggest that the models fine-tuned on our GPT-4 self-instruct data significantly outperformed the Japanese-Alpaca across all three base pre-trained models. Our GPT-4 self-instruct data allowed the LLaMA 13B model to defeat GPT-3.5 (Davinci-003) with a 54.37\% win-rate. The human evaluation exhibits the consistency between GPT-4's assessments and human preference. Our high-quality instruction data and evaluation benchmark have been released here., Comment: COLING 2024. Our code are available here: \href{https://github.com/hitoshizuku7/awesome-Ja-self-instruct}{self-instruct data} and \href{https://github.com/ku-nlp/ja-vicuna-qa-benchmark}{evaluation benchmark}

High-cadence, multiwavelength observations have continuously revealed the diversity of tidal disruption events (TDEs), thus greatly advancing our knowledge and understanding of TDEs. In this work, we conducted an intensive optical-UV and X-ray follow-up campaign of TDE AT2023lli, and found a remarkable month-long bump in its UV/optical light curve nearly two months prior to maximum brightness. The bump represents the longest separation time from the main peak among known TDEs to date. The main UV/optical outburst declines as $t^{-4.10}$, making it one of the fastest decaying optically selected TDEs. Furthermore, we detected sporadic X-ray emission 30 days after the UV/optical peak, accompanied by a reduction in the period of inactivity. It is proposed that the UV/optical bump could be caused by the self-intersection of the stream debris, whereas the primary peak is generated by the reprocessed emission of the accretion process. In addition, our results suggest that episodic X-ray radiation during the initial phase of decline may be due to the patched obscurer surrounding the accretion disk, a phenomenon associated with the inhomogeneous reprocessing process. The double TDE scenario, in which two stars are disrupted in sequence, is also a possible explanation for producing the observed early bump and main peak. We anticipate that the multicolor light curves of TDEs, especially in the very early stages, and the underlying physics can be better understood in the near future with the assistance of dedicated surveys such as the deep high-cadence survey of the 2.5-meter Wide Field Survey Telescope (WFST)., Comment: 14 pages, 8 figures,accepted for publication by ApJL

The oldest stars in the Milky Way (born in the first few billion years) are expected to have a high density in the inner few kpc, spatially overlapping with the Galactic bulge. We use spectroscopic data from the Pristine Inner Galaxy Survey (PIGS) to study the dynamical properties of ancient, metal-poor inner Galaxy stars. We compute distances using StarHorse, and orbital properties in a barred Galactic potential. With this paper, we release the spectroscopic AAT/PIGS catalogue (13 235 stars). We find that most PIGS stars have orbits typical for a pressure-supported population. The fraction of stars confined to the inner Galaxy decreases with decreasing metallicity, but many very metal-poor stars (VMP, [Fe/H] < -2.0) stay confined (~ 60% stay within 5 kpc). The azimuthal velocity v$_\phi$ also decreases between [Fe/H] = -1.0 and -2.0, but is constant for VMP stars (at ~ 40 km/s). The carbon-enhanced metal-poor (CEMP) stars in PIGS appear to have similar orbital properties compared to normal VMP stars. Our results suggest a possible transition between two spheroidal components - a more metal-rich, more concentrated, faster rotating component, and a more metal-poor, more extended and slower/non-rotating component. We propose that the former may be connected to pre-disc in-situ stars (or those born in large building blocks), whereas the latter may be dominated by contributions from smaller galaxies. This is an exciting era where large metal-poor samples, such as in this work (as well as upcoming surveys, e.g., 4MOST), shed light on the earliest evolution of our Galaxy., Comment: Accepted to MNRAS, new: Figures 4, 8 and 9 and Table 2 - Figure 9 shows two inner halo components

While large language models (LLMs) like GPT-4 have recently demonstrated astonishing zero-shot capabilities in general domain tasks, they often generate content with hallucinations in specific domains such as Chinese law, hindering their application in these areas. This is typically due to the absence of training data that encompasses such a specific domain, preventing GPT-4 from acquiring in-domain knowledge. A pressing challenge is that it's not plausible to continue training LLMs of such scale on in-domain data. This paper introduces a simple and effective domain adaptation framework for GPT-4 by reformulating generation as an \textbf{adapt-retrieve-revise} process. The initial step is to \textbf{adapt} an affordable 7B LLM to the target domain by continuing learning on in-domain data. When solving a task, we leverage the adapted LLM to generate a draft answer given a task query. Then, the draft answer will be used to \textbf{retrieve} supporting evidence candidates from an external in-domain knowledge base. Finally, the draft answer and retrieved evidence are concatenated into a whole prompt to let GPT-4 assess the evidence and \textbf{revise} the draft answer to generate the final answer. Our proposal combines the advantages of the efficiency of adapting a smaller 7B model with the evidence-assessing capability of GPT-4 and effectively prevents GPT-4 from generating hallucinatory content. In the zero-shot setting of four Chinese legal tasks, our method improves accuracy by 33.3\% compared to the direct generation by GPT-4. When compared to two stronger retrieval-based baselines, our method outperforms them by 15.4\% and 23.9\%. Our code will be released, Comment: Accepted by ACL 2024 Findings

We study 3d quantum gravity with two asymptotically anti-de Sitter regions, in particular, using its relation with coupled Alekseev-Shatashvili theories and Liouville theory. Expressions for the Hartle-Hawking state, thermal $2n$-point functions, torus wormhole correlators and Wheeler-DeWitt wavefunctions in different bases are obtained using the ZZ boundary states in Liouville theory. Exact results in 2d Jackiw-Teitelboim (JT) gravity are uplifted to 3d gravity, with two copies of Liouville theory in 3d gravity playing a similar role as Schwarzian theory in JT gravity. The connection between 3d gravity and the Liouville ZZ boundary states are manifested by viewing BTZ black holes as Maldacena-Maoz wormholes, with the two wormhole boundaries glued along the ZZ boundaries. In this work, we also study the factorization problem of the Hartle-Hawking state in 3d gravity. With the relevant defect operator that imposes the necessary topological constraint for contractibility, the trace formula in gravity is modified in computing the entanglement entropy. This trace matches with the one from von Neumann algebra considerations, further reproducing the Bekenstein-Hawking area formula from entanglement entropy. Lastly, we propose a calculation for off-shell geometrical quantities that are responsible for the ramp behavior in the late time two-point functions, which follows from the understanding of the Liouville FZZT boundary states in the context of 3d gravity, and the identification between Verlinde loop operators in Liouville theory and "baby universe" operators in 3d gravity., Comment: 71 pages+appendices, 33 figures, typos fixed, references added, JHEP accepted version

We consider the problem of defining a microcanonical thermofield double state at fixed energy and angular momentum from the gravitational path integral. A semiclassical approximation to this state is obtained by imposing a mixed boundary condition on an initial time surface. We analyze the corresponding boundary value problem and gravitational action. The overlap of this state with the canonical thermofield double state, which is interpreted as the Hartle-Hawking wavefunction of an eternal black hole in a mini-superspace approximation, is calculated semiclassically. The relevant saddlepoint is a higher-dimensional, rotating generalization of the wedge geometry that has been studied in two-dimensional gravity., Comment: 20 pages

Despite the success of ChatGPT, its performances on most NLP tasks are still well below the supervised baselines. In this work, we looked into the causes, and discovered that its subpar performance was caused by the following factors: (1) token limit in the prompt does not allow for the full utilization of the supervised datasets; (2) mismatch between the generation nature of ChatGPT and NLP tasks; (3) intrinsic pitfalls of LLMs models, e.g., hallucination, overly focus on certain keywords, etc. In this work, we propose a collection of general modules to address these issues, in an attempt to push the limits of ChatGPT on NLP tasks. Our proposed modules include (1) a one-input-multiple-prompts strategy that employs multiple prompts for one input to accommodate more demonstrations; (2) using fine-tuned models for better demonstration retrieval; (3) transforming tasks to formats that are more tailored to the generation nature; (4) employing reasoning strategies that are tailored to addressing the task-specific complexity; (5) the self-verification strategy to address the hallucination issue of LLMs; (6) the paraphrase strategy to improve the robustness of model predictions. We conduct experiments on 21 datasets of 10 representative NLP tasks, including question answering, commonsense reasoning, natural language inference, sentiment analysis, named entity recognition, entity-relation extraction, event extraction, dependency parsing, semantic role labeling, and part-of-speech tagging. Using the proposed assemble of techniques, we are able to significantly boost the performance of ChatGPT on the selected NLP tasks, achieving performances comparable to or better than supervised baselines, or even existing SOTA performances.

In spite of the potential for ground-breaking achievements offered by large language models (LLMs) (e.g., GPT-3), they still lag significantly behind fully-supervised baselines (e.g., fine-tuned BERT) in relation extraction (RE). This is due to the two major shortcomings of LLMs in RE: (1) low relevance regarding entity and relation in retrieved demonstrations for in-context learning; and (2) the strong inclination to wrongly classify NULL examples into other pre-defined labels. In this paper, we propose GPT-RE to bridge the gap between LLMs and fully-supervised baselines. GPT-RE successfully addresses the aforementioned issues by (1) incorporating task-specific entity representations in demonstration retrieval; and (2) enriching the demonstrations with gold label-induced reasoning logic. We evaluate GPT-RE on four widely-used RE datasets, and observe that GPT-RE achieves improvements over not only existing GPT-3 baselines, but also fully-supervised baselines. Specifically, GPT-RE achieves SOTA performances on the Semeval and SciERC datasets, and competitive performances on the TACRED and ACE05 datasets., Comment: Accepted by EMNLP 2023 Main Conference (long paper)

The successful observation of gravitational waves has provided humanity with an additional method to explore the universe, particularly black holes. In this study, we utilize data from LIGO and Virgo gravitational wave observations to test the first law of black hole mechanics, employing two different approaches. We consider the secondary compact object as a perturbation to the primary black hole before the merger, and the remnant black hole as a stationary black hole after the merger. In the pre-merger and post-merger analysis, our results demonstrate consistency with the first law, with an error level of approximate 25% at a 68% credibility level for GW190403 051519. In the full inspiral-merger-ringdown analysis, our results show consistency with the first law of black hole mechanics, with an error level of about 6% at a 68% credibility level and 10% at a 95% credibility level for GW191219 163120. Additionally, we observe that the higher the mass ratio of the gravitational wave source, the more consistent our results are with the first law of black hole mechanics. Overall, our study sheds light on the nature of compact binary coalescence and their implications for black hole mechanics., Comment: 8 pages, 4 figures, 4 tables

We present the results of a spectroscopic survey of the outskirts of 4 globular clusters -- NGC 1261, NGC 4590, NGC 1904, and NGC 1851 -- covering targets within 1 degree from the cluster centres, with 2dF/AAOmega on the Anglo-Australian Telescope (AAT) and FLAMES on the Very Large Telescope (VLT). We extracted chemo-dynamical information for individual stars, from which we estimated the velocity dispersion profile and the rotation of each cluster. The observations are compared to direct $N$-body simulations and appropriate {\sc limepy}/{\sc spes} models for each cluster to interpret the results. In NGC 1851, the detected internal rotation agrees with existing literature, and NGC 1261 shows some rotation signal beyond the truncation radius, likely coming from the escaped stars. We find that the dispersion profiles for both the observations and the simulations for NGC 1261, NGC 1851, and NGC 1904 do not decrease as the {\sc limepy}/{\sc spes} models predict beyond the truncation radius, where the $N$-body simulations show that escaped stars dominate; the dispersion profile of NGC 4590 follows the predictions of the {\sc limepy}/{\sc spes} models, though the data do not effectively extend beyond the truncation radius. The increasing/flat dispersion profiles in the outskirts of NGC 1261, NGC 1851 and NGC 1904, are reproduced by the simulations. Hence, the increasing/flat dispersion profiles of the clusters in question can be explained by the tidal interaction with the Galaxy without introducing dark matter., Comment: 17 pages, 10 figures. Accepted by MNRAS. Comments are welcome

Relation extraction (RE) has achieved remarkable progress with the help of pre-trained language models. However, existing RE models are usually incapable of handling two situations: implicit expressions and long-tail relation types, caused by language complexity and data sparsity. In this paper, we introduce a simple enhancement of RE using $k$ nearest neighbors ($k$NN-RE). $k$NN-RE allows the model to consult training relations at test time through a nearest-neighbor search and provides a simple yet effective means to tackle the two issues above. Additionally, we observe that $k$NN-RE serves as an effective way to leverage distant supervision (DS) data for RE. Experimental results show that the proposed $k$NN-RE achieves state-of-the-art performances on a variety of supervised RE datasets, i.e., ACE05, SciERC, and Wiki80, along with outperforming the best model to date on the i2b2 and Wiki80 datasets in the setting of allowing using DS. Our code and models are available at: https://github.com/YukinoWan/kNN-RE., Comment: EMNLP 2022 (short paper)

A novel data-driven method of modal analysis for complex flow dynamics, termed as reduced-order variational mode decomposition (RVMD), has been proposed, combining the idea of the separation of variables and a state-of-the-art nonstationary signal-processing technique -- variational mode decomposition. It enables a low-redundant adaptive extraction of coherent structures in statistically nonstationary flows, with its modes computed by solving an elaborate optimization problem using the block coordinate descent algorithm. Discussion on the intrinsic relations between RVMD and some classic modal decomposition methods demonstrates that RVMD can be reduced into proper orthogonal decomposition (POD) or discrete Fourier transform (DFT) at particular parameter settings. The significant advantages of RVMD for performing time-frequency analysis are highlighted by a signal-processing analogous categorization of the widely-used modal decomposition techniques. It is also confirmed that the combination of RVMD and the Hilbert spectral analysis provides a physically intuitive way to explore the space-time-frequency characteristics of transient dynamics. Finally, all the appealing features of RVMD mentioned above are well verified via two canonical flow problems: the transient cylinder wake and the rectangular turbulent supersonic screeching jet.

To solve Math Word Problems, human students leverage diverse reasoning logic that reaches different possible equation solutions. However, the mainstream sequence-to-sequence approach of automatic solvers aims to decode a fixed solution equation supervised by human annotation. In this paper, we propose a controlled equation generation solver by leveraging a set of control codes to guide the model to consider certain reasoning logic and decode the corresponding equations expressions transformed from the human reference. The empirical results suggest that our method universally improves the performance on single-unknown (Math23K) and multiple-unknown (DRAW1K, HMWP) benchmarks, with substantial improvements up to 13.2% accuracy on the challenging multiple-unknown datasets., Comment: AACL 2022 short paper

Contrastive pre-training on distant supervision has shown remarkable effectiveness in improving supervised relation extraction tasks. However, the existing methods ignore the intrinsic noise of distant supervision during the pre-training stage. In this paper, we propose a weighted contrastive learning method by leveraging the supervised data to estimate the reliability of pre-training instances and explicitly reduce the effect of noise. Experimental results on three supervised datasets demonstrate the advantages of our proposed weighted contrastive learning approach compared to two state-of-the-art non-weighted baselines.Our code and models are available at: https://github.com/YukinoWan/WCL, Comment: EACL 2023 (Findings)

Word alignment has proven to benefit many-to-many neural machine translation (NMT). However, high-quality ground-truth bilingual dictionaries were used for pre-editing in previous methods, which are unavailable for most language pairs. Meanwhile, the contrastive objective can implicitly utilize automatically learned word alignment, which has not been explored in many-to-many NMT. This work proposes a word-level contrastive objective to leverage word alignments for many-to-many NMT. Empirical results show that this leads to 0.8 BLEU gains for several language pairs. Analyses reveal that in many-to-many NMT, the encoder's sentence retrieval performance highly correlates with the translation quality, which explains when the proposed method impacts translation. This motivates future exploration for many-to-many NMT to improve the encoder's sentence retrieval performance., Comment: NAACL 2022 findings

We present a comprehensive metallicity analysis of the Sagittarius dwarf spheroidal galaxy (Sgr dSph) using $Pristine\,CaHK$ photometry. We base our member selection on $Gaia$ EDR3 astrometry applying a magnitude limit at $G_{0} = 17.3$, and our population study on the metallicity-sensitive photometry from the $Pristine$ Inner Galaxy Survey (PIGS). Working with photometric metallicities instead of spectroscopic metallicities allows us to cover an unprecedented large area ($\sim 100$ square degrees) of the dwarf galaxy, and to study the spatial distribution of its members as function of metallicity with little selection effects. Our study compares the spatial distributions of a metal-poor population of 9719 stars with [Fe/H] $< -1.3$ and a metal rich one of 30115 stars with [Fe/H] $> -1.0$. The photometric Sgr sample also allows us to assemble the largest sample of 1150 very metal-poor Sgr candidates ([Fe/H] $< -2.0$). By investigating and fitting the spatial properties of the metal-rich and metal-poor population, we find a negative metallicity gradient which extends up to 12 degrees from the Sgr center (or $\sim 5.5$ kpc at the distance of Sgr), the limit of our footprint. We conclude that the relative number of metal-poor stars increases in the outer areas of the galaxy, while the central region is dominated by metal-rich stars. These finding suggest an outside-in formation process and are an indication of the extended formation history of Sgr, which has been affected by the tidal interaction between Sgr and the Milky Way., Comment: The paper was accepted on 20 September by MNRAS

Stellar streams are excellent probes of the underlying gravitational potential in which they evolve. In this work, we fit dynamical models to five streams in the Southern Galactic hemisphere, combining observations from the Southern Stellar Stream Spectroscopic Survey (${S}^5$), Gaia EDR3, and the Dark Energy Survey (DES), to measure the mass of the Large Magellanic Cloud (LMC). With an ensemble of streams, we find a mass of the LMC ranging from 14 to $19 \times 10^{10}\ \mathrm{M}_{\odot}$, probed over a range of closest approach times and distances. With the most constraining stream (Orphan-Chenab), we measure an LMC mass of $18.8^{+ 3.5}_{- 4.0} \times 10^{10}\ \mathrm{M}_{\odot}$, probed at a closest approach time of 310 Myr and a closest approach distance of 25.4 kpc. This mass is compatible with previous measurements, showing that a consistent picture is emerging of the LMC's influence on structures in the Milky Way. Using this sample of streams, we find that the LMC's effect depends on the relative orientation of the stream and LMC at their point of closest approach. To better understand this, we present a simple model based on the impulse approximation and we show that the LMC's effect depends both on the magnitude of the velocity kick imparted to the stream and the direction of this kick., Comment: 23 pages, 10 figures, submitted to ApJ

The most metal-deficient stars hold important clues about the early build-up and chemical evolution of the Milky Way, and carbon-enhanced metal-poor (CEMP) stars are of special interest. However, little is known about CEMP stars in the Galactic bulge. In this paper, we use the large spectroscopic sample of metal-poor stars from the Pristine Inner Galaxy Survey (PIGS) to identify CEMP stars ([C/Fe] > +0.7) in the bulge region and to derive a CEMP fraction. We identify 96 new CEMP stars in the inner Galaxy, of which 62 are very metal-poor ([Fe/H] < -2.0); this is more than a ten-fold increase compared to the seven previously known bulge CEMP stars. The cumulative fraction of CEMP stars in PIGS is $42^{\,+14\,}_{\,-13} \%$ for stars with [Fe/H] < -3.0, and decreases to $16^{\,+3\,}_{\,-3} \%$ for [Fe/H] < -2.5 and $5.7^{\,+0.6\,}_{\,-0.5} \%$ for [Fe/H] < -2.0. The PIGS inner Galaxy CEMP fraction for [Fe/H] < -3.0 is consistent with the halo fraction found in the literature, but at higher metallicities the PIGS fraction is substantially lower. While this can partly be attributed to a photometric selection bias, such bias is unlikely to fully explain the low CEMP fraction at higher metallicities. Considering the typical carbon excesses and metallicity ranges for halo CEMP-s and CEMP-no stars, our results point to a possible deficiency of both CEMP-s and CEMP-no stars (especially the more metal-rich) in the inner Galaxy. The former is potentially related to a difference in the binary fraction, whereas the latter may be the result of a fast chemical enrichment in the early building blocks of the inner Galaxy., Comment: 15 pages, 9 figures, accepted for publication in MNRAS

We compute the albedo (or reflectivity) of electromagnetic waves off the electron-positron Hawking plasma that surrounds the horizon of a Quantum Black Hole. We adopt the "modified firewall conjecture" for fuzzballs [arXiv:hep-th/0502050,arXiv:1711.01617], where we consider significant electromagnetic interaction around the horizon. While prior work has treated this problem as an electron-photon scattering process, we find that the incoming quanta interact collectively with the fermionic excitations of the Hawking plasma at low energies. We derive this via two different methods: one using relativistic plasma dispersion relation, and another using the one-loop correction to photon propagator. Both methods find that the reflectivity of long wavelength photons off the Hawking plasma is significant, contrary to previous claims. This leads to the enhancement of the electromagnetic albedo for frequencies comparable to the Hawking temperature of black hole horizons in vacuum. We comment on possible observable consequences of this effect., Comment: 18 pages, 5 figures, JHEP accepted version

As part of a chemo-dynamical survey of five nearby globular clusters with 2dF/AAOmega on the AAT, we have obtained kinematic information for the globular cluster NGC3201. Our new observations confirm the presence of a significant velocity gradient across the cluster which can almost entirely be explained by the high proper motion of the cluster. After subtracting the contribution of this perspective rotation, we found a remaining rotation signal with an amplitude of $\sim1\ km/s$ around a different axis to what we expect from the tidal tails and the potential escapers, suggesting that this rotation is internal and can be a remnant of its formation process. At the outer part, we found a rotational signal that is likely a result from potential escapers. The proper motion dispersion at large radii reported by Bianchini et al. has previously been attributed to dark matter. Here we show that the LOS dispersion between 0.5-1 Jacobi radius is lower, yet above the predictions from an N-body model of NGC3201 that we ran for this study. Based on the simulation, we find that potential escapers cannot fully explain the observed velocity dispersion. We also estimate the effect on the velocity dispersion of different amounts of stellar-mass black holes and unbound stars from the tidal tails with varying escape rates and find that these effects cannot explain the difference between the LOS dispersion and the N-body model. Given the recent discovery of tidal tail stars at large distances from the cluster, a dark matter halo is an unlikely explanation. We show that the effect of binary stars, which is not included in the N-body model, is important and can explain part of the difference in dispersion. We speculate that the remaining difference must be the result of effects not included in the N-body model, such as initial cluster rotation, velocity anisotropy and Galactic substructure., Comment: 14 pages, 13 figures, accepted by MNRAS