Exploring the Relationship Between Model Architecture and In-Context Learning Ability

What is the relationship between model architecture and the ability to perform in-context learning? In this empirical study, we take the first steps towards answering this question. In particular, we evaluate fifteen model architectures across a suite of synthetic in-context learning tasks. The selected architectures represent a broad range of paradigms, including recurrent and convolution-based neural networks, transformers, and emerging attention alternatives. We discover that all considered architectures can perform in-context learning under certain conditions. However, contemporary architectures are found to be the best performing, especially as task complexity grows. Additionally, our follow-up experiments delve into various factors that influence in-context learning. We observe varied sensitivities among architectures with respect to hyperparameter settings. Our study of training dynamics reveals that certain architectures exhibit a smooth, progressive learning trajectory, while others demonstrate periods of stagnation followed by abrupt mastery of the task. Finally, and somewhat surprisingly, we find that several emerging attention alternatives are more robust in-context learners than transformers; since such approaches have constant-sized memory footprints at inference time, this result opens the future possibility of scaling up in-context learning to vastly larger numbers of in-context examples.

Explainable Knowledge Distillation for On-device Chest X-Ray Classification

Automated multi-label chest X-rays (CXR) image classification has achieved substantial progress in clinical diagnosis via utilizing sophisticated deep learning approaches. However, most deep models have high computational demands, which makes them less feasible for compact devices with low computational requirements. To overcome this problem, we propose a knowledge distillation (KD) strategy to create the compact deep learning model for the real-time multi-label CXR image classification. We study different alternatives of CNNs and Transforms as the teacher to distill the knowledge to a smaller student. Then, we employed explainable artificial intelligence (XAI) to provide the visual explanation for the model decision improved by the KD. Our results on three benchmark CXR datasets show that our KD strategy provides the improved performance on the compact student model, thus being the feasible choice for many limited hardware platforms. For instance, when using DenseNet161 as the teacher network, EEEA-Net-C2 achieved an AUC of 83.7%, 87.1%, and 88.7% on the ChestX-ray14, CheXpert, and PadChest datasets, respectively, with fewer parameters of 4.7 million and computational cost of 0.3 billion FLOPS.

NEAR: Named Entity and Attribute Recognition of clinical concepts

Named Entity Recognition (NER) or the extraction of concepts from clinical text is the task of identifying entities in text and slotting them into categories such as problems, treatments, tests, clinical departments, occurrences (such as admission and discharge) and others. NER forms a critical component of processing and leveraging unstructured data from Electronic Health Records (EHR). While identifying the spans and categories of concepts is itself a challenging task, these entities could also have attributes such as negation that pivot their meanings implied to the consumers of the named entities. There has been little research dedicated to identifying the entities and their qualifying attributes together. This research hopes to contribute to the area of detecting entities and their corresponding attributes by modelling the NER task as a supervised, multi-label tagging problem with each of the attributes assigned tagging sequence labels. In this paper, we propose 3 architectures to achieve this multi-label entity tagging: BiLSTM n-CRF, BiLSTM-CRF-Smax-TF and BiLSTM n-CRF-TF. We evaluate these methods on the 2010 i2b2/VA and the i2b2 2012 shared task datasets. Our different models obtain best NER F1 scores of 0. 894 and 0.808 on the i2b2 2010/VA and i2b2 2012 respectively. The highest span based micro-averaged F1 polarity scores obtained were 0.832 and 0.836 on the i2b2 2010/VA and i2b2 2012 datasets respectively, and the highest macro-averaged F1 polarity scores obtained were 0.924 and 0.888 respectively. The modality studies conducted on i2b2 2012 dataset revealed high scores of 0.818 and 0.501 for span based micro-averaged F1 and macro-averaged F1 respectively.

CHIEF: Clustering with Higher-order Motifs in Big Networks

Clustering a group of vertices in networks facilitates applications across different domains, such as social computing and Internet of Things. However, challenges arises for clustering networks with increased scale. This paper proposes a solution which consists of two motif clustering techniques: standard acceleration CHIEF-ST and approximate acceleration CHIEF-AP. Both algorithms first find the maximal k-edge-connected subgraphs within the target networks to lower the network scale, then employ higher-order motifs in clustering. In the first procedure, we propose to lower the network scale by optimizing the network structure with maximal k-edge-connected subgraphs. For CHIEF-ST, we illustrate that all target motifs will be kept after this procedure when the minimum node degree of the target motif is equal or greater than k. For CHIEF-AP, we prove that the eigenvalues of the adjacency matrix and the Laplacian matrix are relatively stable after this step. That is, CHIEF-ST has no influence on motif clustering, whereas CHIEF-AP introduces limited yet acceptable impact. In the second procedure, we employ higher-order motifs, i.e., heterogeneous four-node motifs clustering in higher-order dense networks. The contributions of CHIEF are two-fold: (1) improved efficiency of motif clustering for big networks; (2) verification of higher-order motif significance. The proposed solutions are found to outperform baseline approaches according to experiments on real and synthetic networks, which demonstrates CHIEF's strength in large network analysis. Meanwhile, higher-order motifs are proved to perform better than traditional triangle motifs in clustering.

Web of Scholars: A Scholar Knowledge Graph

In this work, we demonstrate a novel system, namely Web of Scholars, which integrates state-of-the-art mining techniques to search, mine, and visualize complex networks behind scholars in the field of Computer Science. Relying on the knowledge graph, it provides services for fast, accurate, and intelligent semantic querying as well as powerful recommendations. In addition, in order to realize information sharing, it provides an open API to be served as the underlying architecture for advanced functions. Web of Scholars takes advantage of knowledge graph, which means that it will be able to access more knowledge if more search exist. It can be served as a useful and interoperable tool for scholars to conduct in-depth analysis within Science of Science.

Heterogeneous Graph Learning for Explainable Recommendation over Academic Networks

With the explosive growth of new graduates with research degrees every year, unprecedented challenges arise for early-career researchers to find a job at a suitable institution. This study aims to understand the behavior of academic job transition and hence recommend suitable institutions for PhD graduates. Specifically, we design a deep learning model to predict the career move of early-career researchers and provide suggestions. The design is built on top of scholarly/academic networks, which contains abundant information about scientific collaboration among scholars and institutions. We construct a heterogeneous scholarly network to facilitate the exploring of the behavior of career moves and the recommendation of institutions for scholars. We devise an unsupervised learning model called HAI (Heterogeneous graph Attention InfoMax) which aggregates attention mechanism and mutual information for institution recommendation. Moreover, we propose scholar attention and meta-path attention to discover the hidden relationships between several meta-paths. With these mechanisms, HAI provides ordered recommendations with explainability. We evaluate HAI upon a real-world dataset against baseline methods. Experimental results verify the effectiveness and efficiency of our approach.

Masked Measurement Prediction: Learning to Jointly Predict Quantities and Units from Textual Context

Physical measurements constitute a large portion of numbers in academic papers, engineering reports, and web tables. Current benchmarks fall short of properly evaluating numeracy of pretrained language models on measurements, hindering research on developing new methods and applying them to numerical tasks. To that end, we introduce a novel task, Masked Measurement Prediction (MMP), where a model learns to reconstruct a number together with its associated unit given masked text. MMP is useful for both training new numerically informed models as well as evaluating numeracy of existing systems. In order to address this task, we introduce a new Generative Masked Measurement (GeMM) model that jointly learns to predict numbers along with their units. We perform fine-grained analyses comparing our model with various ablations and baselines. We use linear probing of traditional pretrained transformer models (RoBERTa) to show that they significantly underperform jointly trained number-unit models, highlighting the difficulty of this new task and the benefits of our proposed pretraining approach. We hope this framework accelerates the progress towards building more robust numerical reasoning systems in the future.

Deep Video Anomaly Detection: Opportunities and Challenges

Anomaly detection is a popular and vital task in various research contexts, which has been studied for several decades. To ensure the safety of people's lives and assets, video surveillance has been widely deployed in various public spaces, such as crossroads, elevators, hospitals, banks, and even in private homes. Deep learning has shown its capacity in a number of domains, ranging from acoustics, images, to natural language processing. However, it is non-trivial to devise intelligent video anomaly detection systems cause anomalies significantly differ from each other in different application scenarios. There are numerous advantages if such intelligent systems could be realised in our daily lives, such as saving human resources in a large degree, reducing financial burden on the government, and identifying the anomalous behaviours timely and accurately. Recently, many studies on extending deep learning models for solving anomaly detection problems have emerged, resulting in beneficial advances in deep video anomaly detection techniques. In this paper, we present a comprehensive review of deep learning-based methods to detect the video anomalies from a new perspective. Specifically, we summarise the opportunities and challenges of deep learning models on video anomaly detection tasks, respectively. We put forth several potential future research directions of intelligent video anomaly detection system in various application domains. Moreover, we summarise the characteristics and technical problems in current deep learning methods for video anomaly detection.

EEEA-Net: An Early Exit Evolutionary Neural Architecture Search

The goals of this research were to search for Convolutional Neural Network (CNN) architectures, suitable for an on-device processor with limited computing resources, performing at substantially lower Network Architecture Search (NAS) costs. A new algorithm entitled an Early Exit Population Initialisation (EE-PI) for Evolutionary Algorithm (EA) was developed to achieve both goals. The EE-PI reduces the total number of parameters in the search process by filtering the models with fewer parameters than the maximum threshold. It will look for a new model to replace those models with parameters more than the threshold. Thereby, reducing the number of parameters, memory usage for model storage and processing time while maintaining the same performance or accuracy. The search time was reduced to 0.52 GPU day. This is a huge and significant achievement compared to the NAS of 4 GPU days achieved using NSGA-Net, 3,150 GPU days by the AmoebaNet model, and the 2,000 GPU days by the NASNet model. As well, Early Exit Evolutionary Algorithm networks (EEEA-Nets) yield network architectures with minimal error and computational cost suitable for a given dataset as a class of network algorithms. Using EEEA-Net on CIFAR-10, CIFAR-100, and ImageNet datasets, our experiments showed that EEEA-Net achieved the lowest error rate among state-of-the-art NAS models, with 2.46% for CIFAR-10, 15.02% for CIFAR-100, and 23.8% for ImageNet dataset. Further, we implemented this image recognition architecture for other tasks, such as object detection, semantic segmentation, and keypoint detection tasks, and, in our experiments, EEEA-Net-C2 outperformed MobileNet-V3 on all of these various tasks. (The algorithm code is available at https://github.com/chakkritte/EEEA-Net).

OFFER: A Motif Dimensional Framework for Network Representation Learning

Aiming at better representing multivariate relationships, this paper investigates a motif dimensional framework for higher-order graph learning. The graph learning effectiveness can be improved through OFFER. The proposed framework mainly aims at accelerating and improving higher-order graph learning results. We apply the acceleration procedure from the dimensional of network motifs. Specifically, the refined degree for nodes and edges are conducted in two stages: (1) employ motif degree of nodes to refine the adjacency matrix of the network; and (2) employ motif degree of edges to refine the transition probability matrix in the learning process. In order to assess the efficiency of the proposed framework, four popular network representation algorithms are modified and examined. By evaluating the performance of OFFER, both link prediction results and clustering results demonstrate that the graph representation learning algorithms enhanced with OFFER consistently outperform the original algorithms with higher efficiency.