Search

The Knowledge Graph Completion~(KGC) task aims to infer the missing entity from an incomplete triple. Existing embedding-based methods rely solely on triples in the KG, which is vulnerable to specious relation patterns and long-tail entities. On the other hand, text-based methods struggle with the semantic gap between KG triples and natural language. Apart from triples, entity contexts (e.g., labels, descriptions, aliases) also play a significant role in augmenting KGs. To address these limitations, we propose KGR3, a context-enriched framework for KGC. KGR3 is composed of three modules. Firstly, the Retrieval module gathers supporting triples from the KG, collects plausible candidate answers from a base embedding model, and retrieves context for each related entity. Then, the Reasoning module employs a large language model to generate potential answers for each query triple. Finally, the Re-ranking module combines candidate answers from the two modules mentioned above, and fine-tunes an LLM to provide the best answer. Extensive experiments on widely used datasets demonstrate that KGR3 consistently improves various KGC methods. Specifically, the best variant of KGR3 achieves absolute Hits@1 improvements of 12.3% and 5.6% on the FB15k237 and WN18RR datasets.

Inductive knowledge graph completion (KGC) aims to predict missing triples with unseen entities. Recent works focus on modeling reasoning paths between the head and tail entity as direct supporting evidence. However, these methods depend heavily on the existence and quality of reasoning paths, which limits their general applicability in different scenarios. In addition, we observe that latent type constraints and neighboring facts inherent in KGs are also vital in inferring missing triples. To effectively utilize all useful information in KGs, we introduce CATS, a novel context-aware inductive KGC solution. With sufficient guidance from proper prompts and supervised fine-tuning, CATS activates the strong semantic understanding and reasoning capabilities of large language models to assess the existence of query triples, which consist of two modules. First, the type-aware reasoning module evaluates whether the candidate entity matches the latent entity type as required by the query relation. Then, the subgraph reasoning module selects relevant reasoning paths and neighboring facts, and evaluates their correlation to the query triple. Experiment results on three widely used datasets demonstrate that CATS significantly outperforms state-of-the-art methods in 16 out of 18 transductive, inductive, and few-shot settings with an average absolute MRR improvement of 7.2%.

Selecting a sample generation scheme from multiple text-based generative models is typically addressed by choosing the model that maximizes an averaged evaluation score. However, this score-based selection overlooks the possibility that different models achieve the best generation performance for different types of text prompts. An online identification of the best generation model for various input prompts can reduce the costs associated with querying sub-optimal models. In this work, we explore the possibility of varying rankings of text-based generative models for different text prompts and propose an online learning framework to predict the best data generation model for a given input prompt. The proposed framework adapts the kernelized contextual bandit (CB) methodology to a CB setting with shared context variables across arms, utilizing the generated data to update a kernel-based function that predicts which model will achieve the highest score for unseen text prompts. Additionally, we apply random Fourier features (RFF) to the kernelized CB algorithm to accelerate the online learning process and establish a $\widetilde{\mathcal{O}}(\sqrt{T})$ regret bound for the proposed RFF-based CB algorithm over T iterations. Our numerical experiments on real and simulated text-to-image and image-to-text generative models show RFF-UCB performs successfully in identifying the best generation model across different sample types.

In galaxies with significant ongoing star formation there is an impressively tight correlation between total infrared luminosity (L$_{TIR}$) and H$\alpha$ luminosity (L$_{H\alpha}$), when H$\alpha$ is properly corrected for stellar absorption and dust attenuation. This long-standing result gives confidence that both measurements provide accurate estimates of a galaxy's star formation rate (SFR), despite their differing origins. To test the extent to which this holds in galaxies with lower specific SFR (sSFR=SFR/Mgal, where Mgal is the stellar mass), we combine optical spectroscopy from the Sloan Digital Sky Survey (SDSS) with multi-wavelength (FUV to FIR) photometric observations from the Galaxy And Mass Assembly survey (GAMA). We find that L$_{TIR}$/L$_{H\alpha}$increases steadily with decreasing H$\alpha$ equivalent width (W$_{H\alpha}$, a proxy for sSFR), indicating that both luminosities cannot provide a valid measurement of SFR in galaxies below the canonical star-forming sequence. For both `retired galaxies' and `post-starburst galaxies', L$_{TIR}$/L$_{H\alpha}$ can be up to a factor of 30 larger than for star-forming galaxies. The smooth change in L$_{TIR}$/L$_{H\alpha}$, irrespective of star formation history, ionisation or heating source, dust temperature or other properties, suggests that the value of L$_{TIR}$/L$_{H\alpha}$ is given by the balance between star-forming regions and ambient interstellar medium contributing to both L$_{TIR}$ and L$_{H\alpha}$. While L$_{H\alpha}$ can only be used to estimate the SFR for galaxies with W$_{H\alpha}$ > 3A (sSFR $\gtrsim 10^{-11.5}$/yr), we argue that the mid- and far-infrared can only be used to estimate the SFR of galaxies on the star-forming sequence, and in particular only for galaxies with W$_{H\alpha}$ >10 A (sSFR $\gtrsim 10^{-10.5}$/yr). We find no evidence for dust obscured star-formation in post-starburst galaxies., Comment: Submitted to the Open Journal of Astrophysics. Comments welcome. 14 pages, 4 figures

Existing frameworks for evaluating and comparing generative models typically target an offline setting, where the evaluator has access to full batches of data produced by the models. However, in many practical scenarios, the goal is to identify the best model using the fewest generated samples to minimize the costs of querying data from the models. Such an online comparison is challenging with current offline assessment methods. In this work, we propose an online evaluation framework to find the generative model that maximizes a standard assessment score among a group of available models. Our method uses an optimism-based multi-armed bandit framework to identify the model producing data with the highest evaluation score, quantifying the quality and diversity of generated data. Specifically, we study the online assessment of generative models based on the Fr\'echet Inception Distance (FID) and Inception Score (IS) metrics and propose the FID-UCB and IS-UCB algorithms leveraging the upper confidence bound approach in online learning. We prove sub-linear regret bounds for these algorithms and present numerical results on standard image datasets, demonstrating their effectiveness in identifying the score-maximizing generative model., Comment: arXiv version

The Knowledge Graph Entity Typing (KGET) task aims to predict missing type annotations for entities in knowledge graphs. Recent works only utilize the \textit{\textbf{structural knowledge}} in the local neighborhood of entities, disregarding \textit{\textbf{semantic knowledge}} in the textual representations of entities, relations, and types that are also crucial for type inference. Additionally, we observe that the interaction between semantic and structural knowledge can be utilized to address the false-negative problem. In this paper, we propose a novel \textbf{\underline{S}}emantic and \textbf{\underline{S}}tructure-aware KG \textbf{\underline{E}}ntity \textbf{\underline{T}}yping~{(SSET)} framework, which is composed of three modules. First, the \textit{Semantic Knowledge Encoding} module encodes factual knowledge in the KG with a Masked Entity Typing task. Then, the \textit{Structural Knowledge Aggregation} module aggregates knowledge from the multi-hop neighborhood of entities to infer missing types. Finally, the \textit{Unsupervised Type Re-ranking} module utilizes the inference results from the two models above to generate type predictions that are robust to false-negative samples. Extensive experiments show that SSET significantly outperforms existing state-of-the-art methods., Comment: Accepted in NAACL2024 main

Reinforcement learning (RL) problems where the learner attempts to infer an unobserved reward from some feedback variables have been studied in several recent papers. The setting of Interaction-Grounded Learning (IGL) is an example of such feedback-based RL tasks where the learner optimizes the return by inferring latent binary rewards from the interaction with the environment. In the IGL setting, a relevant assumption used in the RL literature is that the feedback variable $Y$ is conditionally independent of the context-action $(X,A)$ given the latent reward $R$. In this work, we propose Variational Information-based IGL (VI-IGL) as an information-theoretic method to enforce the conditional independence assumption in the IGL-based RL problem. The VI-IGL framework learns a reward decoder using an information-based objective based on the conditional mutual information (MI) between $(X,A)$ and $Y$. To estimate and optimize the information-based terms for the continuous random variables in the RL problem, VI-IGL leverages the variational representation of mutual information to obtain a min-max optimization problem. Also, we extend the VI-IGL framework to general $f$-Information measures leading to the generalized $f$-VI-IGL framework for the IGL-based RL problems. We present numerical results on several reinforcement learning settings indicating an improved performance compared to the existing IGL-based RL algorithm.

We study risk-sensitive Reinforcement Learning (RL), where we aim to maximize the Conditional Value at Risk (CVaR) with a fixed risk tolerance $\tau$. Prior theoretical work studying risk-sensitive RL focuses on the tabular Markov Decision Processes (MDPs) setting. To extend CVaR RL to settings where state space is large, function approximation must be deployed. We study CVaR RL in low-rank MDPs with nonlinear function approximation. Low-rank MDPs assume the underlying transition kernel admits a low-rank decomposition, but unlike prior linear models, low-rank MDPs do not assume the feature or state-action representation is known. We propose a novel Upper Confidence Bound (UCB) bonus-driven algorithm to carefully balance the interplay between exploration, exploitation, and representation learning in CVaR RL. We prove that our algorithm achieves a sample complexity of $\tilde{O}\left(\frac{H^7 A^2 d^4}{\tau^2 \epsilon^2}\right)$ to yield an $\epsilon$-optimal CVaR, where $H$ is the length of each episode, $A$ is the capacity of action space, and $d$ is the dimension of representations. Computational-wise, we design a novel discretized Least-Squares Value Iteration (LSVI) algorithm for the CVaR objective as the planning oracle and show that we can find the near-optimal policy in a polynomial running time with a Maximum Likelihood Estimation oracle. To our knowledge, this is the first provably efficient CVaR RL algorithm in low-rank MDPs., Comment: The first three authors contribute equally and are ordered randomly

We use the stellar fossil record to constrain the stellar metallicity evolution and star-formation histories of the post-starburst (PSB) regions within 45 local post-starburst galaxies from the MaNGA survey. The direct measurement of the regions' stellar metallicity evolution is achieved by a new two-step metallicity model that allows for stellar metallicity to change at the peak of the starburst. We also employ a Gaussian process noise model that accounts for correlated errors introduced by the observational data reduction or inaccuracies in the models. We find that a majority of PSB regions (69% at $>1\sigma$ significance) increased in stellar metallicity during the recent starburst, with an average increase of 0.8 dex and a standard deviation of 0.4 dex. A much smaller fraction of PSBs are found to have remained constant (22%) or declined in metallicity (9%, average decrease 0.4 dex, standard deviation 0.3 dex). The pre-burst metallicities of the PSB galaxies are in good agreement with the mass-metallicity relation of local star-forming galaxies. These results are consistent with hydrodynamic simulations, which suggest that mergers between gas-rich galaxies are the primary formation mechanism of local PSBs, and rapid metal recycling during the starburst outweighs the impact of dilution by any gas inflows. The final mass-weighted metallicities of the PSB galaxies are consistent with the mass-metallicity relation of local passive galaxies. Our results suggest that rapid quenching following a merger-driven starburst is entirely consistent with the observed gap between the stellar mass-metallicity relations of local star-forming and passive galaxies., Comment: 19+5 pages, 8+4 figures, Published in MNRAS

Fixed points represent equilibrium states, stability, and solutions to a range of problems. It has been an active field of research. In this paper, we provide an overview of the main branches of fixed point theory. We discuss the key results and applications.

Image super-resolution is a common task on mobile and IoT devices, where one often needs to upscale and enhance low-resolution images and video frames. While numerous solutions have been proposed for this problem in the past, they are usually not compatible with low-power mobile NPUs having many computational and memory constraints. In this Mobile AI challenge, we address this problem and propose the participants to design an efficient quantized image super-resolution solution that can demonstrate a real-time performance on mobile NPUs. The participants were provided with the DIV2K dataset and trained INT8 models to do a high-quality 3X image upscaling. The runtime of all models was evaluated on the Synaptics VS680 Smart Home board with a dedicated edge NPU capable of accelerating quantized neural networks. All proposed solutions are fully compatible with the above NPU, demonstrating an up to 60 FPS rate when reconstructing Full HD resolution images. A detailed description of all models developed in the challenge is provided in this paper., Comment: arXiv admin note: text overlap with arXiv:2105.07825, arXiv:2105.08826, arXiv:2211.04470, arXiv:2211.03885, arXiv:2211.05256

Post-starburst galaxies (PSBs) have recently and rapidly quenched their star-formation, thus they are an important way to understand how galaxies transition from star-forming late-types to quiescent early-types. The recent discovery of large cold gas reservoirs in PSBs calls into question the theory that galaxies must lose their gas to become quiescent. Optical Integral Field Spectroscopy (IFS) surveys have revealed two classes of PSBs: central PSBs with central quenching regions and ring PSBs with quenching in their outskirts. We analyze a sample of 13 nearby (z < 0.1) PSBs with spatially resolved optical IFS data from the Mapping Nearby Galaxies at Apache Point Observatory (MaNGA) survey and matched resolution Atacama Large (sub-)Millimeter Array (ALMA) observations of $^{12}$CO(1-0). Disturbed stellar kinematics in 7/13 of our PSBs and centrally concentrated molecular gas is consistent with a recent merger for most of our sample. In galaxies without merger evidence, alternate processes may funnel gas inwards and suppress star-formation, which may include outflows, stellar bars, and minor mergers or interactions. The star-formation efficiencies of the post-starburst regions in nearly half our galaxies are suppressed while the gas fractions are consistent with star-forming galaxies. AGN feedback may drive this stabilization, and we observe AGN-consistent emission in the centers of 5/13 galaxies. Finally, our central and ring PSBs have similar properties except the ionized and molecular gas in central PSBs is more disturbed. Overall, the molecular gas in our PSBs tends to be compact and highly disturbed, resulting in concentrated gas reservoirs unable to form stars efficiently., Comment: Accepted for publication in The Astrophysical Journal

Galaxy mergers trigger both star formation and accretion onto the central supermassive black hole. As a result of subsequent energetic feedback processes, it has long been proposed that star formation may be promptly extinguished in galaxy merger remnants. However, this prediction of widespread, rapid quenching in late stage mergers has been recently called into question with modern simulations and has never been tested observationally. Here we perform the first empirical assessment of the long-predicted end phase in the merger sequence. Based on a sample of ~500 post-mergers identified from the Ultraviolet Near Infrared Optical Northern Survey (UNIONS), we show that the frequency of post-merger galaxies that have rapidly shutdown their star formation following a previous starburst is 30-60 times higher than expected from a control sample of non-merging galaxies. No such excess is found in a sample of close galaxy pairs, demonstrating that mergers can indeed lead to a rapid halt to star formation, but that this process only manifests after coalescence., Comment: Accepted for publication in MNRAS Letters

The rapid advances in technology over the last decade have significantly altered the nature of engineering knowledge and skills required in the modern industries. In response to the changing professional requirements, engineering institutions have updated their curriculum and pedagogical practices. However, most of the changes in the curriculum have been focused on the core engineering courses without much consideration for the auxiliary courses in mathematics and sciences. In this paper, we aim to propose a new, augmented mathematics curriculum aimed at meeting the requirements of the modern, technology-based engineering workplace. The proposed updates require minimal resources and can be seamlessly integrated into the existing curriculum., Comment: Published in: 2022 Advances in Science and Engineering Technology International Conferences (IEEE)

Human activity recognition (HAR) through wearable devices has received much interest due to its numerous applications in fitness tracking, wellness screening, and supported living. As a result, we have seen a great deal of work in this field. Traditional deep learning (DL) has set a state of the art performance for HAR domain. However, it ignores the data's structure and the association between consecutive time stamps. To address this constraint, we offer an approach based on Graph Neural Networks (GNNs) for structuring the input representation and exploiting the relations among the samples. However, even when using a simple graph convolution network to eliminate this shortage, there are still several limiting factors, such as inter-class activities issues, skewed class distribution, and a lack of consideration for sensor data priority, all of which harm the HAR model's performance. To improve the current HAR model's performance, we investigate novel possibilities within the framework of graph structure to achieve highly discriminated and rich activity features. We propose a model for (1) time-series-graph module that converts raw data from HAR dataset into graphs; (2) Graph Convolutional Neural Networks (GCNs) to discover local dependencies and correlations between neighboring nodes; and (3) self-attention GNN encoder to identify sensors interactions and data priorities. To the best of our knowledge, this is the first work for HAR, which introduces a GNN-based approach that incorporates both the GCN and the attention mechanism. By employing a uniform evaluation method, our framework significantly improves the performance on hospital patient's activities dataset comparatively considered other state of the art baseline methods.

Although learning has found wide application in multi-agent systems, its effects on the temporal evolution of a system are far from understood. This paper focuses on the dynamics of Q-learning in large-scale multi-agent systems modeled as population games. We revisit the replicator equation model for Q-learning dynamics and observe that this model is inappropriate for our concerned setting. Motivated by this, we develop a new formal model, which bears a formal connection with the continuity equation in physics. We show that our model always accurately describes the Q-learning dynamics in population games across different initial settings of MASs and game configurations. We also show that our model can be applied to different exploration mechanisms, describe the mean dynamics, and be extended to Q-learning in 2-player and n-player games. Last but not least, we show that our model can provide insights into algorithm parameters and facilitate parameter tuning., Comment: the 21st International Conference on Autonomous Agents and Multiagent Systems(AAMAS 2022)

Multi-agent reinforcement learning (MARL) can model many real world applications. However, many MARL approaches rely on epsilon greedy for exploration, which may discourage visiting advantageous states in hard scenarios. In this paper, we propose a new approach QMIX(SEG) for tackling MARL. It makes use of the value function factorization method QMIX to train per-agent policies and a novel Semantic Epsilon Greedy (SEG) exploration strategy. SEG is a simple extension to the conventional epsilon greedy exploration strategy, yet it is experimentally shown to greatly improve the performance of MARL. We first cluster actions into groups of actions with similar effects and then use the groups in a bi-level epsilon greedy exploration hierarchy for action selection. We argue that SEG facilitates semantic exploration by exploring in the space of groups of actions, which have richer semantic meanings than atomic actions. Experiments show that QMIX(SEG) largely outperforms QMIX and leads to strong performance competitive with current state-of-the-art MARL approaches on the StarCraft Multi-Agent Challenge (SMAC) benchmark.

This paper documents the seventeenth data release (DR17) from the Sloan Digital Sky Surveys; the fifth and final release from the fourth phase (SDSS-IV). DR17 contains the complete release of the Mapping Nearby Galaxies at Apache Point Observatory (MaNGA) survey, which reached its goal of surveying over 10,000 nearby galaxies. The complete release of the MaNGA Stellar Library (MaStar) accompanies this data, providing observations of almost 30,000 stars through the MaNGA instrument during bright time. DR17 also contains the complete release of the Apache Point Observatory Galactic Evolution Experiment 2 (APOGEE-2) survey which publicly releases infra-red spectra of over 650,000 stars. The main sample from the Extended Baryon Oscillation Spectroscopic Survey (eBOSS), as well as the sub-survey Time Domain Spectroscopic Survey (TDSS) data were fully released in DR16. New single-fiber optical spectroscopy released in DR17 is from the SPectroscipic IDentification of ERosita Survey (SPIDERS) sub-survey and the eBOSS-RM program. Along with the primary data sets, DR17 includes 25 new or updated Value Added Catalogs (VACs). This paper concludes the release of SDSS-IV survey data. SDSS continues into its fifth phase with observations already underway for the Milky Way Mapper (MWM), Local Volume Mapper (LVM) and Black Hole Mapper (BHM) surveys., Comment: 40 pages, 8 figures, 6 tables. In press at ApJSS (arxiv v2 corrects some minor typos and updates references)

To aid the understanding of the non-linear relationship between galaxy properties and predicted spectral energy distributions (SED), we present a new interactive graphical user interface (GUI) tool pipes_vis based on Bagpipes \citep{arXiv:1712.04452,arXiv:1903.11082}. It allows for real-time manipulation of a model galaxy's star formation history, dust and other relevant properties through sliders and text boxes, with each change's effect on the predicted SED reflected instantaneously. We hope the tool will assist in building intuition about what affects the SED of galaxies, potentially helping to speed up fitting stages such as prior construction, and aid in undergraduate and graduate teaching. pipes_vis is available online (pipes_vis is maintained and documented online at https://github.com/HinLeung622/pipes_vis, or version 0.4.1 is archived in Zenodo and also available for installation through pip install pipes_vis)., Comment: 3 pages, 1 figure, Research Notes of the AAS accepted

With recent advances in modelling stars using high-precision asteroseismology, the systematic effects associated with our assumptions of stellar helium abundance ($Y$) and the mixing-length theory parameter ($\alpha_\mathrm{MLT}$) are becoming more important. We apply a new method to improve the inference of stellar parameters for a sample of Kepler dwarfs and subgiants across a narrow mass range ($0.8 < M < 1.2\,\mathrm{M_\odot}$). In this method, we include a statistical treatment of $Y$ and the $\alpha_\mathrm{MLT}$. We develop a hierarchical Bayesian model to encode information about the distribution of $Y$ and $\alpha_\mathrm{MLT}$ in the population, fitting a linear helium enrichment law including an intrinsic spread around this relation and normal distribution in $\alpha_\mathrm{MLT}$. We test various levels of pooling parameters, with and without solar data as a calibrator. When including the Sun as a star, we find the gradient for the enrichment law, $\Delta Y / \Delta Z = 1.05^{+0.28}_{-0.25}$ and the mean $\alpha_\mathrm{MLT}$ in the population, $\mu_\alpha = 1.90^{+0.10}_{-0.09}$. While accounting for the uncertainty in $Y$ and $\alpha_\mathrm{MLT}$, we are still able to report statistical uncertainties of 2.5 per cent in mass, 1.2 per cent in radius, and 12 per cent in age. Our method can also be applied to larger samples which will lead to improved constraints on both the population level inference and the star-by-star fundamental parameters., Comment: 21 pages, 15 figures, 12 tables. Accepted for publication in Monthly Notices of the Royal Astronomical Society

In teacher-student framework, a more experienced agent (teacher) helps accelerate the learning of another agent (student) by suggesting actions to take in certain states. In cooperative multiagent reinforcement learning (MARL), where agents need to cooperate with one another, a student may fail to cooperate well with others even by following the teachers' suggested actions, as the polices of all agents are ever changing before convergence. When the number of times that agents communicate with one another is limited (i.e., there is budget constraint), the advising strategy that uses actions as advices may not be good enough. We propose a partaker-sharer advising framework (PSAF) for cooperative MARL agents learning with budget constraint. In PSAF, each Q-learner can decide when to ask for Q-values and share its Q-values. We perform experiments in three typical multiagent learning problems. Evaluation results show that our approach PSAF outperforms existing advising methods under both unlimited and limited budget, and we give an analysis of the impact of advising actions and sharing Q-values on agents' learning., Comment: 31 pages, 16 figures, submitted to ACM Transactions on Autonomous and Adaptive Systems (TAAS)