Entity retrieval--retrieving information about entity mentions in a query--is a key step in open-domain tasks, such as question answering or fact checking. However, state-of-the-art entity retrievers struggle to retrieve rare entities for ambiguous mentions due to biases towards popular entities. Incorporating knowledge graph types during training could help overcome popularity biases, but there are several challenges: (1) existing type-based retrieval methods require mention boundaries as input, but open-domain tasks run on unstructured text, (2) type-based methods should not compromise overall performance, and (3) type-based methods should be robust to noisy and missing types. In this work, we introduce TABi, a method to jointly train bi-encoders on knowledge graph types and unstructured text for entity retrieval for open-domain tasks. TABi leverages a type-enforced contrastive loss to encourage entities and queries of similar types to be close in the embedding space. TABi improves retrieval of rare entities on the Ambiguous Entity Retrieval (AmbER) sets, while maintaining strong overall retrieval performance on open-domain tasks in the KILT benchmark compared to state-of-the-art retrievers. TABi is also robust to incomplete type systems, improving rare entity retrieval over baselines with only 5% type coverage of the training dataset. We make our code publicly available at https://github.com/HazyResearch/tabi.
In this research we present a quantum circuit for estimating algorithmic information metrics like the universal prior distribution. This accelerates inferring algorithmic structure in data for discovering causal generative models. The computation model is restricted in time and space resources to make it computable in approximating the target metrics. A classical exhaustive enumeration is shown for a few examples. The precise quantum circuit design that allows executing a superposition of automata is presented. As a use-case, an application framework for experimenting on DNA sequences for meta-biology is proposed. To our knowledge, this is the first time approximating algorithmic information is implemented for quantum computation. Our implementation on the OpenQL quantum programming language and the QX Simulator is copy-left and can be found on https://github.com/Advanced-Research-Centre/QuBio.
A novel algorithm to detect semantic lines is proposed in this paper. We develop three networks: detection network with mirror attention (D-Net) and comparative ranking and matching networks (R-Net and M-Net). D-Net extracts semantic lines by exploiting rich contextual information. To this end, we design the mirror attention module. Then, through pairwise comparisons of extracted semantic lines, we iteratively select the most semantic line and remove redundant ones overlapping with the selected one. For the pairwise comparisons, we develop R-Net and M-Net in the Siamese architecture. Experiments demonstrate that the proposed algorithm outperforms the conventional semantic line detector significantly. Moreover, we apply the proposed algorithm to detect two important kinds of semantic lines successfully: dominant parallel lines and reflection symmetry axes. Our codes are available at https://github.com/dongkwonjin/Semantic-Line-DRM.
In recent years, hypergraph learning has attracted great attention due to its capacity in representing complex and high-order relationships. However, current neural network approaches designed for hypergraphs are mostly shallow, thus limiting their ability to extract information from high-order neighbors. In this paper, we show both theoretically and empirically, that the performance of hypergraph neural networks does not improve as the number of layers increases, which is known as the over-smoothing problem. To tackle this issue, we develop a new deep hypergraph convolutional network called Deep-HGCN, which can maintain the heterogeneity of node representation in deep layers. Specifically, we prove that a $k$-layer Deep-HGCN simulates a polynomial filter of order $k$ with arbitrary coefficients, which can relieve the problem of over-smoothing. Experimental results on various datasets demonstrate the superior performance of the proposed model comparing to the state-of-the-art hypergraph learning approaches.
Finding and selecting the most relevant scientific papers from a large number of papers written in a research community is one of the key challenges for researchers these days. As we know, much information around research interest for scholars and academicians belongs to papers they read. Analysis and extracting contextual features from these papers could help us to suggest the most related paper to them. In this paper, we present a multi-task recommendation system (RS) that predicts a paper recommendation and generates its meta-data such as keywords. The system is implemented as a three-stage deep neural network encoder that tries to maps longer sequences of text to an embedding vector and learns simultaneously to predict the recommendation rate for a particular user and the paper's keywords. The motivation behind this approach is that the paper's topics expressed as keywords are a useful predictor of preferences of researchers. To achieve this goal, we use a system combination of RNNs, Highway and Convolutional Neural Networks to train end-to-end a context-aware collaborative matrix. Our application uses Highway networks to train the system very deep, combine the benefits of RNN and CNN to find the most important factor and make latent representation. Highway Networks allow us to enhance the traditional RNN and CNN pipeline by learning more sophisticated semantic structural representations. Using this method we can also overcome the cold start problem and learn latent features over large sequences of text.
Clinical notes in Electronic Health Records (EHR) present rich documented information of patients to inference phenotype for disease diagnosis and study patient characteristics for cohort selection. Unsupervised user embedding aims to encode patients into fixed-length vectors without human supervisions. Medical concepts extracted from the clinical notes contain rich connections between patients and their clinical categories. However, existing unsupervised approaches of user embeddings from clinical notes do not explicitly incorporate medical concepts. In this study, we propose a concept-aware unsupervised user embedding that jointly leverages text documents and medical concepts from two clinical corpora, MIMIC-III and Diabetes. We evaluate user embeddings on both extrinsic and intrinsic tasks, including phenotype classification, in-hospital mortality prediction, patient retrieval, and patient relatedness. Experiments on the two clinical corpora show our approach exceeds unsupervised baselines, and incorporating medical concepts can significantly improve the baseline performance.
Graph Neural Networks (GNNs) have shown promising results in various tasks, among which link prediction is an important one. GNN models usually follow a node-centric message passing procedure that aggregates the neighborhood information to the central node recursively. Following this paradigm, features of nodes are passed through edges without caring about where the nodes are located and which role they played. However, the neglected topological information is shown to be valuable for link prediction tasks. In this paper, we propose Structure Enhanced Graph neural network (SEG) for link prediction. SEG introduces the path labeling method to capture surrounding topological information of target nodes and then incorporates the structure into an ordinary GNN model. By jointly training the structure encoder and deep GNN model, SEG fuses topological structures and node features to take full advantage of graph information. Experiments on the OGB link prediction datasets demonstrate that SEG achieves state-of-the-art results among all three public datasets.
For stochastic models with intractable likelihood functions, approximate Bayesian computation offers a way of approximating the true posterior through repeated comparisons of observations with simulated model outputs in terms of a small set of summary statistics. These statistics need to retain the information that is relevant for constraining the parameters but cancel out the noise. They can thus be seen as thermodynamic state variables, for general stochastic models. For many scientific applications, we need strictly more summary statistics than model parameters to reach a satisfactory approximation of the posterior. Therefore, we propose to use the inner dimension of deep neural network based Autoencoders as summary statistics. To create an incentive for the encoder to encode all the parameter-related information but not the noise, we give the decoder access to explicit or implicit information on the noise that has been used to generate the training data. We validate the approach empirically on two types of stochastic models.
The increased throughput brought by MIMO technology relies on the knowledge of channel state information (CSI) acquired in the base station (BS). To make the CSI feedback overhead affordable for the evolution of MIMO technology (e.g., massive MIMO and ultra-massive MIMO), deep learning (DL) is introduced to deal with the CSI compression task. Based on the separation principle in existing communication systems, DL based CSI compression is used as source coding. However, this separate source-channel coding (SSCC) scheme is inferior to the joint source-channel coding (JSCC) scheme in the finite blocklength regime. In this paper, we propose a deep joint source-channel coding (DJSCC) based framework for the CSI feedback task. In particular, the proposed method can simultaneously learn from the CSI source and the wireless channel. Instead of truncating CSI via Fourier transform in the delay domain in existing methods, we apply non-linear transform networks to compress the CSI. Furthermore, we adopt an SNR adaption mechanism to deal with the wireless channel variations. The extensive experiments demonstrate the validity, adaptability, and generality of the proposed framework.
The success of neural networks on a diverse set of NLP tasks has led researchers to question how much do these networks actually know about natural language. Probes are a natural way of assessing this. When probing, a researcher chooses a linguistic task and trains a supervised model to predict annotation in that linguistic task from the network's learned representations. If the probe does well, the researcher may conclude that the representations encode knowledge related to the task. A commonly held belief is that using simpler models as probes is better; the logic is that such models will identify linguistic structure, but not learn the task itself. We propose an information-theoretic formalization of probing as estimating mutual information that contradicts this received wisdom: one should always select the highest performing probe one can, even if it is more complex, since it will result in a tighter estimate. The empirical portion of our paper focuses on obtaining tight estimates for how much information BERT knows about parts of speech in a set of five typologically diverse languages that are often underrepresented in parsing research, plus English, totaling six languages. We find BERT accounts for only at most 5% more information than traditional, type-based word embeddings.