Sentence insertion is an interesting NLP problem but received insufficient attention. Existing approaches in sentence ordering, text coherence, and question answering are neither suitable nor good enough at solving it. To bridge this gap, we propose InsertGNN, a simple yet effective model that represents the problem as a graph and adopts a hierarchical graph neural network (GNN) to learn the connection between sentences. We evaluate our method in our newly collected TOEFL dataset and further verify its effectiveness on the larger arXiv dataset using cross-domain learning. Extensive experiments demonstrate that InsertGNN outperforms all baselines by a large margin with an accuracy of 70\%, rivaling the average human test scores.

Federated learning (FL) provides an effective machine learning (ML) architecture to protect data privacy in a distributed manner. However, the inevitable network asynchrony, the over-dependence on a central coordinator, and the lack of an open and fair incentive mechanism collectively hinder its further development. We propose \textsc{IronForge}, a new generation of FL framework, that features a Directed Acyclic Graph (DAG)-based data structure and eliminates the need for central coordinators to achieve fully decentralized operations. \textsc{IronForge} runs in a public and open network, and launches a fair incentive mechanism by enabling state consistency in the DAG, so that the system fits in networks where training resources are unevenly distributed. In addition, dedicated defense strategies against prevalent FL attacks on incentive fairness and data privacy are presented to ensure the security of \textsc{IronForge}. Experimental results based on a newly developed testbed FLSim highlight the superiority of \textsc{IronForge} to the existing prevalent FL frameworks under various specifications in performance, fairness, and security. To the best of our knowledge, \textsc{IronForge} is the first secure and fully decentralized FL framework that can be applied in open networks with realistic network and training settings.

Fifteen Puzzle problem is one of the most classical problems that have captivated mathematical enthusiasts for centuries. This is mainly because of the huge size of the state space with approximately 1013 states that have to be explored and several algorithms have been applied to solve the Fifteen Puzzle instances. In this paper, to deal with this large state space, Bidirectional A* (BA*) search algorithm with three heuristics, such as Manhattan distance (MD), linear conflict (LC), and walking distance (WD) has been used to solve the Fifteen Puzzle problems. The three mentioned heuristics will be hybridized in a way that can dramatically reduce the number of generated states by the algorithm. Moreover, all those heuristics require only 25KB of storage but help the algorithm effectively reduce the number of generated states and expand fewer nodes. Our implementation of BA* search can significantly reduce the space complexity, and guarantee either optimal or near-optimal solutions.1

In recent years we have witnessed a renewed interest in machine learning methodologies, especially for deep representation learning, that could overcome basic i.i.d. assumptions and tackle non-stationary environments subject to various distributional shifts or sample selection biases. Within this context, several computational approaches based on architectural priors, regularizers and replay policies have been proposed with different degrees of success depending on the specific scenario in which they were developed and assessed. However, designing comprehensive hybrid solutions that can flexibly and generally be applied with tunable efficiency-effectiveness trade-offs still seems a distant goal. In this paper, we propose "Architect, Regularize and Replay" (ARR), an hybrid generalization of the renowned AR1 algorithm and its variants, that can achieve state-of-the-art results in classic scenarios (e.g. class-incremental learning) but also generalize to arbitrary data streams generated from real-world datasets such as CIFAR-100, CORe50 and ImageNet-1000.

Automated visual story generation aims to produce stories with corresponding illustrations that exhibit coherence, progression, and adherence to characters' emotional development. This work proposes a story generation pipeline to co-create visual stories with the users. The pipeline allows the user to control events and emotions on the generated content. The pipeline includes two parts: narrative and image generation. For narrative generation, the system generates the next sentence using user-specified keywords and emotion labels. For image generation, diffusion models are used to create a visually appealing image corresponding to each generated sentence. Further, object recognition is applied to the generated images to allow objects in these images to be mentioned in future story development.

Dynamic Graph Neural Networks (DGNNs) have been broadly applied in various real-life applications, such as link prediction and pandemic forecast, to capture both static structural information and temporal characteristics from dynamic graphs. Combining both time-dependent and -independent components, DGNNs manifest substantial parallel computation and data reuse potentials, but suffer from severe memory access inefficiency and data transfer overhead under the canonical one-graph-at-a-time training pattern. To tackle the challenges, we propose PiPAD, a $\underline{\textbf{Pi}}pelined$ and $\underline{\textbf{PA}}rallel$ $\underline{\textbf{D}}GNN$ training framework for the end-to-end performance optimization on GPUs. From both the algorithm and runtime level, PiPAD holistically reconstructs the overall training paradigm from the data organization to computation manner. Capable of processing multiple graph snapshots in parallel, PiPAD eliminates the unnecessary data transmission and alleviates memory access inefficiency to improve the overall performance. Our evaluation across various datasets shows PiPAD achieves $1.22\times$-$9.57\times$ speedup over the state-of-the-art DGNN frameworks on three representative models.

The grammatical analysis of texts in any written language typically involves a number of basic processing tasks, such as tokenization, morphological tagging, and dependency parsing. State-of-the-art systems can achieve high accuracy on these tasks for languages with large datasets, but yield poor results for languages which have little to no annotated data. To address this issue for the Tagalog language, we investigate the use of alternative language resources for creating task-specific models in the absence of dependency-annotated Tagalog data. We also explore the use of word embeddings and data augmentation to improve performance when only a small amount of annotated Tagalog data is available. We show that these zero-shot and few-shot approaches yield substantial improvements on grammatical analysis of both in-domain and out-of-domain Tagalog text compared to state-of-the-art supervised baselines.

The diversity of terrestrial vascular plants plays a key role in maintaining the stability and productivity of ecosystems. Monitoring species compositional diversity across large spatial scales is challenging and time consuming. The advanced spectral and spatial specification of the recently launched DESIS (the DLR Earth Sensing Imaging Spectrometer) instrument provides a unique opportunity to test the potential for monitoring plant species diversity with spaceborne hyperspectral data. This study provides a quantitative assessment on the ability of DESIS hyperspectral data for predicting plant species richness in two different habitat types in southeast Australia. Spectral features were first extracted from the DESIS spectra, then regressed against on-ground estimates of plant species richness, with a two-fold cross validation scheme to assess the predictive performance. We tested and compared the effectiveness of Principal Component Analysis (PCA), Canonical Correlation Analysis (CCA), and Partial Least Squares analysis (PLS) for feature extraction, and Kernel Ridge Regression (KRR), Gaussian Process Regression (GPR), Random Forest Regression (RFR) for species richness prediction. The best prediction results were r=0.76 and RMSE=5.89 for the Southern Tablelands region, and r=0.68 and RMSE=5.95 for the Snowy Mountains region. Relative importance analysis for the DESIS spectral bands showed that the red-edge, red, and blue spectral regions were more important for predicting plant species richness than the green bands and the near-infrared bands beyond red-edge. We also found that the DESIS hyperspectral data performed better than Sentinel-2 multispectral data in the prediction of plant species richness. Our results provide a quantitative reference for future studies exploring the potential of spaceborne hyperspectral data for plant biodiversity mapping.

While GPT-2 generates sentences that are remarkably human-like, longer documents can ramble and do not follow human-like writing structure. We study the problem of imposing structure on long-range text. We propose a novel controlled text generation task, sequentially controlled text generation, and identify a dataset, NewsDiscourse as a starting point for this task. We develop a sequential controlled text generation pipeline with generation and editing. We test different degrees of structural awareness and show that, in general, more structural awareness results in higher control-accuracy, grammaticality, coherency and topicality, approaching human-level writing performance.

Combinatorics math problems are often used as a benchmark to test human cognitive and logical problem-solving skills. These problems are concerned with counting the number of solutions that exist in a specific scenario that is sketched in natural language. Humans are adept at solving such problems as they can identify commonly occurring structures in the questions for which a closed-form formula exists for computing the answer. These formulas exploit the exchangeability of objects and symmetries to avoid a brute-force enumeration of all possible solutions. Unfortunately, current AI approaches are still unable to solve combinatorial problems in this way. This paper aims to fill this gap by developing novel AI techniques for representing and solving such problems. It makes the following five contributions. First, we identify a class of combinatorics math problems which traditional lifted counting techniques fail to model or solve efficiently. Second, we propose a novel declarative language for this class of problems. Third, we propose novel lifted solving algorithms bridging probabilistic inference techniques and constraint programming. Fourth, we implement them in a lifted solver that solves efficiently the class of problems under investigation. Finally, we evaluate our contributions on a real-world combinatorics math problems dataset and synthetic benchmarks.