Place recognition is the fundamental module that can assist Simultaneous Localization and Mapping (SLAM) in loop-closure detection and re-localization for long-term navigation. The place recognition community has made astonishing progress over the last $20$ years, and this has attracted widespread research interest and application in multiple fields such as computer vision and robotics. However, few methods have shown promising place recognition performance in complex real-world scenarios, where long-term and large-scale appearance changes usually result in failures. Additionally, there is a lack of an integrated framework amongst the state-of-the-art methods that can handle all of the challenges in place recognition, which include appearance changes, viewpoint differences, robustness to unknown areas, and efficiency in real-world applications. In this work, we survey the state-of-the-art methods that target long-term localization and discuss future directions and opportunities. We start by investigating the formulation of place recognition in long-term autonomy and the major challenges in real-world environments. We then review the recent works in place recognition for different sensor modalities and current strategies for dealing with various place recognition challenges. Finally, we review the existing datasets for long-term localization and introduce our datasets and evaluation API for different approaches. This paper can be a tutorial for researchers new to the place recognition community and those who care about long-term robotics autonomy. We also provide our opinion on the frequently asked question in robotics: Do robots need accurate localization for long-term autonomy? A summary of this work and our datasets and evaluation API is publicly available to the robotics community at: https://github.com/MetaSLAM/GPRS.
Lexical paraphrasing aims at acquiring word-level paraphrases. It is critical for many Natural Language Processing (NLP) applications, such as Question Answering (QA), Information Extraction (IE), and Machine Translation (MT). Since the meaning and usage of a word can vary in distinct contexts, different paraphrases should be acquired according to the contexts. However, most of the existing researches focus on constructing paraphrase corpora, in which little contextual constraints for paraphrase application are imposed. This paper presents a method that automatically acquires context-specific lexical paraphrases. In this method, the obtained paraphrases of a word depend on the specific sentence the word occurs in. Two stages are included, i.e. candidate paraphrase extraction and paraphrase validation, both of which are mainly based on web mining. Evaluations are conducted on a news title corpus and the presented method is compared with a paraphrasing method that exploits a Chinese thesaurus of synonyms -- Tongyi Cilin (Extended) (CilinE for short). Results show that the f-measure of our method (0.4852) is significantly higher than that using CilinE (0.1127). In addition, over 85% of the correct paraphrases derived by our method cannot be found in CilinE, which suggests that our method is effective in acquiring out-of-thesaurus paraphrases.
Defect detection is extremely important to improve the quality of PCB production. Although defect detection using traditional methods has achieved good results, a large number of false detections and missed detections cannot be avoided. In response to solve this problems, we propose a fine-grained defect detection network (FDDNet) model to improve the detection performance of PCB defects. This model increases the dimension of spatial context features in PCB defect detection to fuse multi-scale features, which helps the model to deal with more complex scenes. To facilitate the efficiency of feature fusion, we propose an improved channel attention module to enhance the learning efficiency of the network for effective features. To cooperate with the multiplexing of multi-scale feature maps in the backbone network, we propose a module capable of enhancing image recognition to extract pure shallow information. Finally, the experimental results on the PCB defect dataset show that the proposed method can achieve a mAP50 index of 97.32%.
This paper proposes a novel method to resolve the coverage problem of SMT system. The method generates paraphrases for source-side sentences of the bilingual parallel data, which are then paired with the target-side sentences to generate new parallel data. Within a statistical paraphrase generation framework, we employ an object function, named Sentence Novelty, to select paraphrases which having the most novel information to the bilingual training corpus of the SMT model. Meanwhile, the context is considered via a language model in the source language to ensure the fluency and accuracy of paraphrase substitution. Compared to a state-of-the-art phrase based SMT system (Moses), our method achieves an improvement of 1.66 points in terms of BLEU on a small training corpus which simulates a resource-poor environment, and 1.06 points on a training corpus of medium size.
Promoting the development of "unconventional" physics experiments plays an important role in the implementation of China's "double reduction" education policy, and using it as a lever can promote the continuous improvement of basic physics education.The article believes that under the background of the "double reduction" education policy, "unconventional" Physics experiments ushered in two major opportunities for development: policy support and the boost of a good educational environment; and after clarifying the impact of "unconventional" physics experiments on the "double reduction" education policy, leading education to return to life and improving students' interest in learning, etc.The value of the times in many aspects.The article explains how to promote the guidance of "unconventional" physics experiments to cultivate innovative physics teachers and explore the action logic of ecological physics teaching to ensure the effective implementation of China's "double reduction" education policy and realize the high-quality development of basic physics education.
In this study, 90 finite-element models are used to explore the behaviour of fibre-reinforced polymer (FRP) reinforced joints under combined in-plane bending (IPB) and axial load (AX). The effects of joint geometry, FRP layer count, and AX levels of the chord or brace are considered. Three typical failure modes are observed: chord plastic failure, brace plastic failure, and brace buckling failure. Increasing the number of FRP layers can ensure that failure is chord-related failure in a ductility manner rather than the unexpectedly brace-related brittle failure. Depending on the stress distribution of fibres, FRP reinforcement can restrict the deformation of joints subjected to complex loading patterns. Moreover, added FRP layers efficiently reduce the effect of brace AX on the IPB resistance. Finally, a modified strength equation is established, including the influence of FRP reinforcement, chord AX, and brace AX.
The experimental study of heat transfer and flow characteristics are conducted for water and ethylene-glycol solution flowing in the heat exchanger with rectangular or triangular microscale channels, which have equivalent diameter of 0.55mm, 0.91mm, 1.38mm and 5mm. During experiments, the Reynolds number ranges from 300 to 2500. The experimental results show that: at a fixed Reynolds number, the Nusselt number increases along with increasing equivalent diameter, the Nusselt number for ethylene-glycol solution with larger Prandtl number is greater than that for water, the geometrical configuration of the microscale channels have a significant effect on the heat transfer; the flow friction factors of microscale channels are smaller than that of normal channels, flow characteristics of rectangular channels are evidently better than that of triangular, the flow friction factor decreases with increasing Reynolds number, the experiment also show that the flow friction factor is independent of Prandtl number; The critical Reynolds number at which the flow transiting to turbulent flow is 700∼1200.
Question paraphrasing is critical in many Natural Language Processing (NLP) applications, especially for question reformulation in question answering (QA). However, choosing an appropriate data source and developing effective methods are challenging tasks. In this paper, we propose a method that exploits Encarta logs to automatically identify question paraphrases and extract templates. Questions from Encarta logs are partitioned into small clusters, within which a perceptron classier is used for identifying question paraphrases. Experiments are conducted and the results have shown: (1) Encarta log data is an eligible data source for question paraphrasing and the user clicks in the data are indicative clues for recognizing paraphrases; (2) the supervised method we present is effective, which can evidently outperform the unsupervised method. Besides, the features introduced to identify paraphrases are sound; (3) the obtained question paraphrase templates are quite effective in question reformulation, enhancing the MRR from 0.2761 to 0.4939 with the questions of TREC QA 2003.
Paraphrases are various expressions that convey the same meaning. Research of paraphrasing is critical in many related NLP research areas, such as machine translation (MT), question answering (QA), information retrieval (IR), information extraction (IE), natural language generation (NLG), etc.
This tutorial is intended to provide the attendees with an in-depth look at the identification, generation, application, and evaluation of paraphrases. The tutorial first reviews studies on paraphrase identification (or extraction), which aims to acquire paraphrases from various data sources, such as large-scale web corpora, monolingual parallel corpora, monolingual comparable corpora, bilingual parallel corpora, as well as some other resources.
It then surveys methods on paraphrase generation, in which the MT-based method will be highlighted, while the other kinds of methods, including thesaurus-based, pattern-based, and NLG-based methods, will also be introduced.
We then discuss the applications of paraphrases in related research areas, especially in MT. We will show how paraphrases can help to alleviate data sparseness problem, simplify input sentences, tune parameters, and improve automatic evaluation in statistical MT systems.
The last part of the tutorial is about the evaluation of paraphrases. Till now, no approach has been widely accepted on paraphrase evaluation, which leaves it as an open issue. This tutorial will summarize existing approaches to paraphrase evaluation, which include human evaluation, automatic evaluation, and application-driven evaluation.
The target audience will be NLP researchers, practitioners, and students. But participants do not need prior knowledge of paraphrasing.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)
