Paper Group NANR 21
Cross-Task Knowledge Transfer for Visually-Grounded Navigation. Robust ConvNet Landmark-Based Visual Place Recognition by Optimizing Landmark Matching. Structured Knowledge Distillation for Semantic Segmentation. Words are Vectors, Dependencies are Matrices: Learning Word Embeddings from Dependency Graphs. Attack and Anomaly Detection in IoT Sensor …
Cross-Task Knowledge Transfer for Visually-Grounded Navigation
| Title | Cross-Task Knowledge Transfer for Visually-Grounded Navigation |
| Authors | Devendra Singh Chaplot, Lisa Lee, Ruslan Salakhutdinov, Devi Parikh, Dhruv Batra |
| Abstract | Recent efforts on training visual navigation agents conditioned on language using deep reinforcement learning have been successful in learning policies for two different tasks: learning to follow navigational instructions and embodied question answering. In this paper, we aim to learn a multitask model capable of jointly learning both tasks, and transferring knowledge of words and their grounding in visual objects across tasks. The proposed model uses a novel Dual-Attention unit to disentangle the knowledge of words in the textual representations and visual objects in the visual representations, and align them with each other. This disentangled task-invariant alignment of representations facilitates grounding and knowledge transfer across both tasks. We show that the proposed model outperforms a range of baselines on both tasks in simulated 3D environments. We also show that this disentanglement of representations makes our model modular, interpretable, and allows for zero-shot transfer to instructions containing new words by leveraging object detectors. |
| Tasks | Embodied Question Answering, Question Answering, Transfer Learning, Visual Navigation |
| Published | 2019-05-01 |
| URL | https://openreview.net/forum?id=ByGq7hRqKX |
| https://openreview.net/pdf?id=ByGq7hRqKX | |
| PWC | https://paperswithcode.com/paper/cross-task-knowledge-transfer-for-visually |
| Repo | |
| Framework | |
Robust ConvNet Landmark-Based Visual Place Recognition by Optimizing Landmark Matching
| Title | Robust ConvNet Landmark-Based Visual Place Recognition by Optimizing Landmark Matching |
| Authors | Yaguang Kong (ygkong@hdu.edu.cn) and Wei Liu (weiliu994@gmail.com) |
| Abstract | Visual place recognition (VPR) is a fundamental but challenging problem that has not been solved completely for a long time, especially in a kaleidoscopic environment. Recent advanced works which exploit ConvNet landmarks as a representation of an image for the VPR have demonstrated promising performance under condition and viewpoint changes. In this paper, we propose an improved ConvNet landmark-based VPR with better robustness and higher matching efficiency by extending this method from two aspects. First, we introduce hashing to find global optimal landmark matches for each landmark in the query image to boost the quality of matched landmark pairs. Second, we apply the sequence search for finding the best matches basing on the temporal information attached in both query and reference images. The experiments which conducted on four challengeable benchmark datasets show that our approach significantly enhances the robustness of traditional ConvNet landmark-based VPR and outperforms the state-of-the-art ConvNet feature-based VPR named SeqCNNSLAM. Moreover, our method has higher computing efficiency than previous ConvNet landmark-based VPR. |
| Tasks | Visual Place Recognition |
| Published | 2019-02-27 |
| URL | https://ieeexplore.ieee.org/document/8653820 |
| https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8653820 | |
| PWC | https://paperswithcode.com/paper/robust-convnet-landmark-based-visual-place |
| Repo | |
| Framework | |
Structured Knowledge Distillation for Semantic Segmentation
| Title | Structured Knowledge Distillation for Semantic Segmentation |
| Authors | Yifan Liu, Ke Chen, Chris Liu, Zengchang Qin, Zhenbo Luo, Jingdong Wang |
| Abstract | In this paper, we investigate the issue of knowledge distillation for training compact semantic segmentation networks by making use of cumbersome networks. We start from the straightforward scheme, pixel-wise distillation, which applies the distillation scheme originally introduced for image classification and performs knowledge distillation for each pixel separately. We further propose to distill the structured knowledge from cumbersome networks into compact networks, which is motivated by the fact that semantic segmentation is a structured prediction problem. We study two such structured distillation schemes: (i) pair-wise distillation that distills the pairwise similarities, and (ii) holistic distillation that uses adversarial training to distill holistic knowledge. The effectiveness of our knowledge distillation approaches is demonstrated by extensive experiments on three scene parsing datasets: Cityscapes, Camvid and ADE20K. |
| Tasks | Image Classification, Scene Parsing, Semantic Segmentation, Structured Prediction |
| Published | 2019-06-01 |
| URL | http://openaccess.thecvf.com/content_CVPR_2019/html/Liu_Structured_Knowledge_Distillation_for_Semantic_Segmentation_CVPR_2019_paper.html |
| http://openaccess.thecvf.com/content_CVPR_2019/papers/Liu_Structured_Knowledge_Distillation_for_Semantic_Segmentation_CVPR_2019_paper.pdf | |
| PWC | https://paperswithcode.com/paper/structured-knowledge-distillation-for-1 |
| Repo | |
| Framework | |
Words are Vectors, Dependencies are Matrices: Learning Word Embeddings from Dependency Graphs
| Title | Words are Vectors, Dependencies are Matrices: Learning Word Embeddings from Dependency Graphs |
| Authors | Paula Czarnowska, Guy Emerson, Ann Copestake |
| Abstract | Distributional Semantic Models (DSMs) construct vector representations of word meanings based on their contexts. Typically, the contexts of a word are defined as its closest neighbours, but they can also be retrieved from its syntactic dependency relations. In this work, we propose a new dependency-based DSM. The novelty of our model lies in associating an independent meaning representation, a matrix, with each dependency-label. This allows it to capture specifics of the relations between words and contexts, leading to good performance on both intrinsic and extrinsic evaluation tasks. In addition to that, our model has an inherent ability to represent dependency chains as products of matrices which provides a straightforward way of handling further contexts of a word. |
| Tasks | Learning Word Embeddings, Word Embeddings |
| Published | 2019-05-01 |
| URL | https://www.aclweb.org/anthology/W19-0408/ |
| https://www.aclweb.org/anthology/W19-0408 | |
| PWC | https://paperswithcode.com/paper/words-are-vectors-dependencies-are-matrices |
| Repo | |
| Framework | |
Attack and Anomaly Detection in IoT Sensors in IoT Sites Using Machine Learning Approaches
| Title | Attack and Anomaly Detection in IoT Sensors in IoT Sites Using Machine Learning Approaches |
| Authors | Mahmudul Hasan, Md. Milon Islam, Ishrak Islam, M.M.A. Hashem |
| Abstract | Attack and Anomaly detection in the Internet of Things (IoT) infrastructure is a rising concern in the domain of IoT. With the increased use of IoT infrastructure in every domain, threats and attacks in these infrastructures are also growing commensurately. Denial of Service, Data Type Probing, Malicious Control, Malicious Operation, Scan, Spying and Wrong Setup are such attacks and anomalies which can cause an IoT system failure. In this paper, performances of several machine learning models have been compared to predict attacks and anomalies on the IoT systems accurately. The machine learning (ML) algorithms that have been used here are Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), and Artificial Neural Network (ANN). The evaluation metrics used in the comparison of performance are accuracy, precision, recall, f1 score, and area under the Receiver Operating Characteristic Curve. The system obtained 99.4% test accuracy for Decision Tree, Random Forest, and ANN. Though these techniques have the same accuracy, other metrics prove that Random Forest performs comparatively better. |
| Tasks | Anomaly Detection |
| Published | 2019-05-11 |
| URL | https://doi.org/10.1016/j.iot.2019.100059 |
| https://reader.elsevier.com/reader/sd/pii/S2542660519300241?token=65A572C03AF0035AE87F974E9D4CF7B40ED1C0DD7A45692B31722802DF84A277851FA7E27F0A3247061677A0FFDF91FB | |
| PWC | https://paperswithcode.com/paper/attack-and-anomaly-detection-in-iot-sensors |
| Repo | |
| Framework | |
Controlling Sequence-to-Sequence Models - A Demonstration on Neural-based Acrostic Generator
| Title | Controlling Sequence-to-Sequence Models - A Demonstration on Neural-based Acrostic Generator |
| Authors | Liang-Hsin Shen, Pei-Lun Tai, Chao-Chung Wu, Shou-De Lin |
| Abstract | An acrostic is a form of writing that the first token of each line (or other recurring features in the text) forms a meaningful sequence. In this paper we present a generalized acrostic generation system that can hide certain message in a flexible pattern specified by the users. Different from previous works that focus on rule-based solutions, here we adopt a neural- based sequence-to-sequence model to achieve this goal. Besides acrostic, users are also allowed to specify the rhyme and length of the output sequences. Based on our knowledge, this is the first neural-based natural language generation system that demonstrates the capability of performing micro-level control over output sentences. |
| Tasks | Text Generation |
| Published | 2019-11-01 |
| URL | https://www.aclweb.org/anthology/D19-3008/ |
| https://www.aclweb.org/anthology/D19-3008 | |
| PWC | https://paperswithcode.com/paper/controlling-sequence-to-sequence-models-a |
| Repo | |
| Framework | |
Shifting the Baseline: Single Modality Performance on Visual Navigation & QA
| Title | Shifting the Baseline: Single Modality Performance on Visual Navigation & QA |
| Authors | Jesse Thomason, Daniel Gordon, Yonatan Bisk |
| Abstract | We demonstrate the surprising strength of unimodal baselines in multimodal domains, and make concrete recommendations for best practices in future research. Where existing work often compares against random or majority class baselines, we argue that unimodal approaches better capture and reflect dataset biases and therefore provide an important comparison when assessing the performance of multimodal techniques. We present unimodal ablations on three recent datasets in visual navigation and QA, seeing an up to 29{%} absolute gain in performance over published baselines. |
| Tasks | Visual Navigation |
| Published | 2019-06-01 |
| URL | https://www.aclweb.org/anthology/N19-1197/ |
| https://www.aclweb.org/anthology/N19-1197 | |
| PWC | https://paperswithcode.com/paper/shifting-the-baseline-single-modality-1 |
| Repo | |
| Framework | |
R3 Adversarial Network for Cross Model Face Recognition
| Title | R3 Adversarial Network for Cross Model Face Recognition |
| Authors | Ken Chen, Yichao Wu, Haoyu Qin, Ding Liang, Xuebo Liu, Junjie Yan |
| Abstract | In this paper, we raise a new problem, namely cross model face recognition (CMFR), which has considerable economic and social significance. The core of this problem is to make features extracted from different models comparable. However, the diversity, mainly caused by different application scenarios, frequent version updating, and all sorts of service platforms, obstructs interaction among different models and poses a great challenge. To solve this problem, from the perspective of Bayesian modelling, we propose R3 Adversarial Network (R3AN) which consists of three paths: Reconstruction, Representation and Regression. We also introduce adversarial learning into the reconstruction path for better performance. Comprehensive experiments on public datasets demonstrate the feasibility of interaction among different models with the proposed framework. When updating the gallery, R3AN conducts the feature transformation nearly 10 times faster than ResNet-101. Meanwhile, the transformed feature distribution is very close to that of target model, and its error rate is incredibly reduced by approximately 75% compared with a naive transformation model. Furthermore, we show that face feature can be deciphered into original face image roughly by the reconstruction path, which may give valuable hints for improving the original face recognition models. |
| Tasks | Face Recognition |
| Published | 2019-06-01 |
| URL | http://openaccess.thecvf.com/content_CVPR_2019/html/Chen_R3_Adversarial_Network_for_Cross_Model_Face_Recognition_CVPR_2019_paper.html |
| http://openaccess.thecvf.com/content_CVPR_2019/papers/Chen_R3_Adversarial_Network_for_Cross_Model_Face_Recognition_CVPR_2019_paper.pdf | |
| PWC | https://paperswithcode.com/paper/r3-adversarial-network-for-cross-model-face |
| Repo | |
| Framework | |
Correlated Uncertainty for Learning Dense Correspondences from Noisy Labels
| Title | Correlated Uncertainty for Learning Dense Correspondences from Noisy Labels |
| Authors | Natalia Neverova, David Novotny, Andrea Vedaldi |
| Abstract | Many machine learning methods depend on human supervision to achieve optimal performance. However, in tasks such as DensePose, where the goal is to establish dense visual correspondences between images, the quality of manual annotations is intrinsically limited. We address this issue by augmenting neural network predictors with the ability to output a distribution over labels, thus explicitly and introspectively capturing the aleatoric uncertainty in the annotations. Compared to previous works, we show that correlated error fields arise naturally in applications such as DensePose and these fields can be modeled by deep networks, leading to a better understanding of the annotation errors. We show that these models, by understanding uncertainty better, can solve the original DensePose task more accurately, thus setting the new state-of-the-art accuracy in this benchmark. Finally, we demonstrate the utility of the uncertainty estimates in fusing the predictions of produced by multiple models, resulting in a better and more principled approach to model ensembling which can further improve accuracy. |
| Tasks | |
| Published | 2019-12-01 |
| URL | http://papers.nips.cc/paper/8378-correlated-uncertainty-for-learning-dense-correspondences-from-noisy-labels |
| http://papers.nips.cc/paper/8378-correlated-uncertainty-for-learning-dense-correspondences-from-noisy-labels.pdf | |
| PWC | https://paperswithcode.com/paper/correlated-uncertainty-for-learning-dense |
| Repo | |
| Framework | |
SemEval-2019 Task 12: Toponym Resolution in Scientific Papers
| Title | SemEval-2019 Task 12: Toponym Resolution in Scientific Papers |
| Authors | Davy Weissenbacher, Arjun Magge, Karen O{'}Connor, Matthew Scotch, Gonzalez-Hern, Graciela ez |
| Abstract | We present the SemEval-2019 Task 12 which focuses on toponym resolution in scientific articles. Given an article from PubMed, the task consists of detecting mentions of names of places, or toponyms, and mapping the mentions to their corresponding entries in GeoNames.org, a database of geospatial locations. We proposed three subtasks. In Subtask 1, we asked participants to detect all toponyms in an article. In Subtask 2, given toponym mentions as input, we asked participants to disambiguate them by linking them to entries in GeoNames. In Subtask 3, we asked participants to perform both the detection and the disambiguation steps for all toponyms. A total of 29 teams registered, and 8 teams submitted a system run. We summarize the corpus and the tools created for the challenge. They are freely available at https://competitions.codalab.org/competitions/19948. We also analyze the methods, the results and the errors made by the competing systems with a focus on toponym disambiguation. |
| Tasks | |
| Published | 2019-06-01 |
| URL | https://www.aclweb.org/anthology/S19-2155/ |
| https://www.aclweb.org/anthology/S19-2155 | |
| PWC | https://paperswithcode.com/paper/semeval-2019-task-12-toponym-resolution-in |
| Repo | |
| Framework | |
A Report on the Third VarDial Evaluation Campaign
| Title | A Report on the Third VarDial Evaluation Campaign |
| Authors | Marcos Zampieri, Shervin Malmasi, Yves Scherrer, Tanja Samard{\v{z}}i{'c}, Francis Tyers, Miikka Silfverberg, Natalia Klyueva, Tung-Le Pan, Chu-Ren Huang, Radu Tudor Ionescu, Andrei M. Butnaru, Tommi Jauhiainen |
| Abstract | In this paper, we present the findings of the Third VarDial Evaluation Campaign organized as part of the sixth edition of the workshop on Natural Language Processing (NLP) for Similar Languages, Varieties and Dialects (VarDial), co-located with NAACL 2019. This year, the campaign included five shared tasks, including one task re-run {–} German Dialect Identification (GDI) {–} and four new tasks {–} Cross-lingual Morphological Analysis (CMA), Discriminating between Mainland and Taiwan variation of Mandarin Chinese (DMT), Moldavian vs. Romanian Cross-dialect Topic identification (MRC), and Cuneiform Language Identification (CLI). A total of 22 teams submitted runs across the five shared tasks. After the end of the competition, we received 14 system description papers, which are published in the VarDial workshop proceedings and referred to in this report. |
| Tasks | Language Identification, Morphological Analysis |
| Published | 2019-06-01 |
| URL | https://www.aclweb.org/anthology/W19-1401/ |
| https://www.aclweb.org/anthology/W19-1401 | |
| PWC | https://paperswithcode.com/paper/a-report-on-the-third-vardial-evaluation |
| Repo | |
| Framework | |
SoftRegex: Generating Regex from Natural Language Descriptions using Softened Regex Equivalence
| Title | SoftRegex: Generating Regex from Natural Language Descriptions using Softened Regex Equivalence |
| Authors | Jun-U Park, Sang-Ki Ko, Marco Cognetta, Yo-Sub Han |
| Abstract | We continue the study of generating se-mantically correct regular expressions from natural language descriptions (NL). The current state-of-the-art model SemRegex produces regular expressions from NLs by rewarding the reinforced learning based on the semantic (rather than syntactic) equivalence between two regular expressions. Since the regular expression equivalence problem is PSPACE-complete, we introduce the EQ{_}Reg model for computing the simi-larity of two regular expressions using deep neural networks. Our EQ{_}Reg mod-el essentially softens the equivalence of two regular expressions when used as a reward function. We then propose a new regex generation model, SoftRegex, us-ing the EQ{_}Reg model, and empirically demonstrate that SoftRegex substantially reduces the training time (by a factor of at least 3.6) and produces state-of-the-art results on three benchmark datasets. |
| Tasks | |
| Published | 2019-11-01 |
| URL | https://www.aclweb.org/anthology/D19-1677/ |
| https://www.aclweb.org/anthology/D19-1677 | |
| PWC | https://paperswithcode.com/paper/softregex-generating-regex-from-natural |
| Repo | |
| Framework | |
Combining Data-Intense and Compute-Intense Methods for Fine-Grained Morphological Analyses
| Title | Combining Data-Intense and Compute-Intense Methods for Fine-Grained Morphological Analyses |
| Authors | Petra Steiner |
| Abstract | |
| Tasks | |
| Published | 2019-09-01 |
| URL | https://www.aclweb.org/anthology/W19-8506/ |
| https://www.aclweb.org/anthology/W19-8506 | |
| PWC | https://paperswithcode.com/paper/combining-data-intense-and-compute-intense |
| Repo | |
| Framework | |
Correlation between the gradability of Latin adjectives and the ability to form qualitative abstract nouns
| Title | Correlation between the gradability of Latin adjectives and the ability to form qualitative abstract nouns |
| Authors | Lucie Pultrov{'a} |
| Abstract | |
| Tasks | |
| Published | 2019-09-01 |
| URL | https://www.aclweb.org/anthology/W19-8504/ |
| https://www.aclweb.org/anthology/W19-8504 | |
| PWC | https://paperswithcode.com/paper/correlation-between-the-gradability-of-latin |
| Repo | |
| Framework | |
The Treatment of Word Formation in the LiLa Knowledge Base of Linguistic Resources for Latin
| Title | The Treatment of Word Formation in the LiLa Knowledge Base of Linguistic Resources for Latin |
| Authors | Eleonora Litta, Marco Passarotti, Francesco Mambrini |
| Abstract | |
| Tasks | |
| Published | 2019-09-01 |
| URL | https://www.aclweb.org/anthology/W19-8505/ |
| https://www.aclweb.org/anthology/W19-8505 | |
| PWC | https://paperswithcode.com/paper/the-treatment-of-word-formation-in-the-lila |
| Repo | |
| Framework | |
