_id stringlengths 4 10 | text stringlengths 0 18.4k | title stringlengths 0 8.56k |
|---|---|---|
d257984946 | The backdoor attack, where the adversary uses inputs stamped with triggers (e.g., a patch) to activate pre-planted malicious behaviors, is a severe threat to Deep Neural Network (DNN) models. Trigger inversion is an effective way of identifying backdoor models and understanding embedded adversarial behaviors. A challenge of trigger inversion is that there are many ways of constructing the trigger. Existing methods cannot generalize to various types of triggers by making certain assumptions or attack-specific constraints. The fundamental reason is that existing work does not consider the trigger's design space in their formulation of the inversion problem. This work formally defines and analyzes the triggers injected in different spaces and the inversion problem. Then, it proposes a unified framework to invert backdoor triggers based on the formalization of triggers and the identified inner behaviors of backdoor models from our analysis. Our prototype UNICORN is general and effective in inverting backdoor triggers in DNNs. The code can be found at https://github.com/ RU-System-Software-and-Security/UNICORN. | UNICORN: A UNIFIED BACKDOOR TRIGGER INVER- SION FRAMEWORK |
d222124941 | Thompson Sampling (TS) is one of the most effective algorithms for solving contextual multi-armed bandit problems. In this paper, we propose a new algorithm, called Neural Thompson Sampling, which adapts deep neural networks for both exploration and exploitation. At the core of our algorithm is a novel posterior distribution of the reward, where its mean is the neural network approximator, and its variance is built upon the neural tangent features of the corresponding neural network. We prove that, provided the underlying reward function is bounded, the proposed algorithm is guaranteed to achieve a cumulative regret of O(T 1/2 ), which matches the regret of other contextual bandit algorithms in terms of total round number T . Experimental comparisons with other benchmark bandit algorithms on various data sets corroborate our theory. arXiv:2010.00827v2 [cs.LG] 30 Dec 2021 | Published as a conference paper at ICLR 2021 NEURAL THOMPSON SAMPLING |
d246863709 | Quantization of deep neural networks (DNN) has been proven effective for compressing and accelerating DNN models. Data-free quantization (DFQ) is a promising approach without the original datasets under privacy-sensitive and confidential scenarios. However, current DFQ solutions degrade accuracy, need synthetic data to calibrate networks, and are time-consuming and costly. This paper proposes an on-the-fly DFQ framework with sub-second quantization time, called SQuant, which can quantize networks on inference-only devices with low computation and memory requirements. With the theoretical analysis of the second-order information of DNN task loss, we decompose and approximate the Hessian-based optimization objective into three diagonal sub-items, which have different areas corresponding to three dimensions of weight tensor: element-wise, kernel-wise, and output channel-wise. Then, we progressively compose sub-items and propose a novel data-free optimization objective in the discrete domain, minimizing Constrained Absolute Sum of Error (or CASE in short), which surprisingly does not need any dataset and is even not aware of network architecture. We also design an efficient algorithm without back-propagation to further reduce the computation complexity of the objective solver. Finally, without fine-tuning and synthetic datasets, SQuant accelerates the data-free quantization process to a sub-second level with > 30% accuracy improvement over the existing data-free post-training quantization works, with the evaluated models under 4-bit quantization. We have open-sourced the SQuant framework 1 . * Jingwen Leng and Minyi Guo are corresponding authors of this paper. . Post-training 4-bit quantization of convolution networks for rapid-deployment. arXiv preprint arXiv:1810.05723, 2018. -v2: Hessian aware trace-weighted quantization of neural networks. arXiv preprint arXiv:1911.03852, 2019a. . Hawq: Hessian aware quantization of neural networks with mixed-precision. In . Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity. arXiv preprint arXiv:2101.03961, 2021.Trevor Gale, Matei Zaharia, Cliff Young, and Erich Elsen. Sparse gpu kernels for deep learning. . Accelerating sparse dnn models without hardware-support via tile-wise sparsity. In efficiency and flexibility for dnn acceleration via temporal gpu-systolic array integration. | Published as a conference paper at ICLR 2022 SQUANT: ON-THE-FLY DATA-FREE QUANTIZATION VIA DIAGONAL HESSIAN APPROXIMATION |
d53221030 | We consider a simple and overarching representation for permutation-invariant functions of sequences (or set functions). Our approach, which we call Janossy pooling, expresses a permutation-invariant function as the average of a permutation-sensitive function applied to all reorderings of the input sequence. This allows us to leverage the rich and mature literature on permutation-sensitive functions to construct novel and flexible permutation-invariant functions. If carried out naively, Janossy pooling can be computationally prohibitive. To allow computational tractability, we consider three kinds of approximations: canonical orderings of sequences, functions with k-order interactions, and stochastic optimization algorithms with random permutations. Our framework unifies a variety of existing work in the literature, and suggests possible modeling and algorithmic extensions. We explore a few in our experiments, which demonstrate improved performance over current state-of-the-art methods. | JANOSSY POOLING: LEARNING DEEP PERMUTATION- INVARIANT FUNCTIONS FOR VARIABLE-SIZE INPUTS |
d235422602 | The discovery of the disentanglement properties of the latent space in GANs motivated a lot of research to find the semantically meaningful directions on it. In this paper, we suggest that the disentanglement property is closely related to the geometry of the latent space. In this regard, we propose an unsupervised method for finding the semantic-factorizing directions on the intermediate latent space of GANs based on the local geometry. Intuitively, our proposed method, called Local Basis, finds the principal variation of the latent space in the neighborhood of the base latent variable. Experimental results show that the local principal variation corresponds to the semantic factorization and traversing along it provides strong robustness to image traversal. Moreover, we suggest an explanation for the limited success in finding the global traversal directions in the latent space, especially W-space of StyleGAN2. We show that W-space is warped globally by comparing the local geometry, discovered from Local Basis, through the metric on Grassmannian Manifold. The global warpage implies that the latent space is not well-aligned globally and therefore the global traversal directions are bound to show limited success on it. * Equal contribution † Corresponding author arXiv:2106.06959v5 [cs.CV] 25 Jun 2022Published as a conference paper at ICLR 2022 method that describes the local disentanglement property and an evaluation scheme for the global disentanglement property from the collected local information. | DO NOT ESCAPE FROM THE MANIFOLD: DISCOVER- ING THE LOCAL COORDINATES ON THE LATENT SPACE OF GANS |
d248006342 | We present a new framework AMOS that pretrains text encoders with an Adversarial learning curriculum via a Mixture Of Signals from multiple auxiliary generators. Following ELECTRA-style pretraining, the main encoder is trained as a discriminator to detect replaced tokens generated by auxiliary masked language models (MLMs). Different from ELECTRA which trains one MLM as the generator, we jointly train multiple MLMs of different sizes to provide training signals at various levels of difficulty. To push the discriminator to learn better with challenging replaced tokens, we learn mixture weights over the auxiliary MLMs' outputs to maximize the discriminator loss by backpropagating the gradient from the discriminator via Gumbel-Softmax. For better pretraining efficiency, we propose a way to assemble multiple MLMs into one unified auxiliary model. AMOS outperforms ELECTRA and recent state-of-the-art pretrained models by about 1 point on the GLUE benchmark for BERT base-sized models.Recent studies revealed that the key to the ELECTRA's success is its new learning dynamics(Xu et al., 2020;Meng et al., 2021). By pretraining the auxiliary model jointly with the main Transformer, an implicit learning curriculum is formed: The noise produced by the auxiliary generator becomes more and more plausible during pretraining, posing greater challenges for the discriminator, which has to overcome the difficulty by reasoning more deeply using the contexts. This leads to significantly improved sample efficiency and effectiveness of ELECTRA-style pretrained models(Clark et al., 2020;Chi et al., 2021;Meng et al., 2021).On the other hand, such a training dynamic also introduced new challenges in search of the optimal pretraining setting. First, the configurations of the auxiliary generator-its depth, width, and masking fraction-require costly trial-and-error pretraining runs. At the same time, they also significantly impact the discriminator's downstream task performance: A weak auxiliary model does not generate hard enough pretraining signal to push the discriminator, but a too strong one can confuse the discriminator and worsen its downstream task performance(Clark et al., 2020;Meng et al., 2021). Second, the side-by-side training of the two models forms a pseudo "GAN-style"(Goodfellow et al., 2014)curriculum which causes difficulty to improve or scale: Previous attempts to make the generator and discriminator learning more interactive (e.g., training the generator to maximize the discriminator loss as in actual GAN frameworks) resulted in downgraded performance(Clark et al., 2020). | Published as a conference paper at ICLR 2022 PRETRAINING TEXT ENCODERS WITH ADVERSARIAL MIXTURE OF TRAINING SIGNAL GENERATORS |
d247058667 | Indiscriminate data poisoning attacks are quite effective against supervised learning. However, not much is known about their impact on unsupervised contrastive learning (CL). This paper is the first to consider indiscriminate poisoning attacks of contrastive learning. We propose Contrastive Poisoning (CP), the first effective such attack on CL. We empirically show that Contrastive Poisoning, not only drastically reduces the performance of CL algorithms, but also attacks supervised learning models, making it the most generalizable indiscriminate poisoning attack. We also show that CL algorithms with a momentum encoder are more robust to indiscriminate poisoning, and propose a new countermeasure based on matrix completion. | INDISCRIMINATE POISONING ATTACKS ON UNSUPER- VISED CONTRASTIVE LEARNING |
d259088773 | Finetuning large language models inflates the costs of NLU applications and remains the bottleneck of development cycles. Recent works in computer vision use data pruning to reduce training time. Pruned data selection with static methods is based on a score calculated for each training example prior to finetuning, which involves important computational overhead. Moreover, the score may not necessarily be representative of sample importance throughout the entire training duration. We propose to address these issues with a refined version of dynamic data pruning, a curriculum which periodically scores and discards unimportant examples during finetuning. Our method leverages an EL2N metric that we extend to the joint intent and slot classification task, and an initial finetuning phase on the full train set. Our results on the GLUE benchmark and four joint NLU datasets show a better timeaccuracy trade-off compared to static methods. Our method preserves full accuracy while training on 50% of the data points and reduces computational times by up to 41%. If we tolerate instead a minor drop of accuracy of 1%, we can prune 80% of the training examples for a reduction in finetuning time reaching 66%. . 2020. Self-paced learning for neural machine translation. In | NLU on Data Diets: Dynamic Data Subset Selection for NLP Classification Tasks |
d263875066 | In politics, neologisms are frequently invented for partisan objectives. For example, "undocumented workers" and "illegal aliens" refer to the same group of people (i.e., they have the same denotation), but they carry clearly different connotations. Examples like these have traditionally posed a challenge to referencebased semantic theories and led to increasing acceptance of alternative theories (e.g., Two-Factor Semantics) among philosophers and cognitive scientists. In NLP, however, popular pretrained models encode both denotation and connotation as one entangled representation. In this study, we propose an adversarial neural network that decomposes a pretrained representation as independent denotation and connotation representations. For intrinsic interpretability, we show that words with the same denotation but different connotations (e.g., "immigrants" vs. "aliens", "estate tax" vs. "death tax") move closer to each other in denotation space while moving further apart in connotation space. For extrinsic application, we train an information retrieval system with our disentangled representations and show that the denotation vectors improve the viewpoint diversity of document rankings. | Do "Undocumented Workers" == "Illegal Aliens"? Differentiating Denotation and Connotation in Vector Spaces |
d264038749 | L'utilisation généralisée de documents numériques non sécurisés par les entreprises et les administrations comme pièces justificatives les rend vulnérables à la falsification.En outre, les logiciels de retouche d'images et les possibilités qu'ils offrent compliquent les tâches de la détection de fraude d'images numériques.Néanmoins, la recherche dans ce domaine se heurte au manque de données réalistes accessibles au public.Dans cet article, nous proposons un nouveau jeu de données pour la détection des faux tickets contenant 988 images numérisées de tickets et leurs transcriptions, provenant du jeu de données SROIE (scanned receipts OCR and information extraction).163 images et leurs transcriptions ont subi des modifications frauduleuses réalistes et ont été annotées.Nous décrivons en détail le jeu de données, les falsifications et leurs annotations et fournissons deux baselines (basées sur l'image et le texte) sur la tâche de détection de la fraude. | Jeu de données de tickets de caisse pour la détection de fraude documentaire |
d1520275 | In state-of-the-art Neural Machine Translation (NMT), an attention mechanism is used during decoding to enhance the translation. At every step, the decoder uses this mechanism to focus on different parts of the source sentence to gather the most useful information before outputting its target word. Recently, the effectiveness of the attention mechanism has also been explored for multimodal tasks, where it becomes possible to focus both on sentence parts and image regions that they describe. In this paper, we compare several attention mechanism on the multimodal translation task (English, image → German) and evaluate the ability of the model to make use of images to improve translation. We surpass state-of-the-art scores on the Multi30k data set, we nevertheless identify and report different misbehavior of the machine while translating. | An empirical study on the effectiveness of images in Multimodal Neural Machine Translation |
d263877582 | Stochastic approaches to natural language processing have often been preferred to rule-based approaches because of their robustness and their automatic training capabilities. This was the case for part-of-speech tagging until Brill showed how state-of-the-art part-of-speech tagging can be achieved with a rule-based tagger by inferring rules from a training corpus. However, current implementations of the rule-based tagger run more slowly than previous approaches. In this paper, we present a finite-state tagger, inspired by the rule-based tagger, that operates in optimal time in the sense that the time to assign tags to a sentence corresponds to the time required to follow a single path in a deterministic finite-state machine. This result is achieved by encoding the application of the rules found in the tagger as a nondeterministic finite-state transducer and then turning it into a deterministic transducer. The resulting deterministic transducer yields a part-of-speech tagger whose speed is dominated by the access time of mass storage devices. We then generalize the techniques to the class of transformation-based systems. | Deterministic Part-of-Speech Tagging with Finite-State Transducers |
d1131746 | We describe the use of clinical data present in the medical record to determine the relevance of research evidence from literature databases. We studied the effect of using automated knowledge approaches as compared to physician's selection of articles, when using a traditional information retrieval system. Three methods were evaluated. The first method identified terms and their semantics and relationships in the patient's record to build a map of the record, which was represented in conceptual graph notation. This approach was applied to data in an individual's medical record and used to score citations retrieved using a graph matching algorithm. The second method identified associations between terms in the medical record, assigning them semantic types and weights based on the co-occurrence of these associations in citations of biomedical literature. The method was applied to data in an individual's medical record and used to score citations. The last method combined the first two. The results showed that physicians agreed better with each other than with the automated methods. However, we found a significant positive relation between physicians' selection of abstracts and two of the methods. We believe the results encourage the use of clinical data to determine the relevance of medical literature to the care of individual patients. | Analyzing the Semantics of Patient Data to Rank Records of Literature Retrieval |
d1076 | In this paper, we present a method of estimating referents of demonstrative pronouns, personal pronouns, and zero pronouns in Japanese sentences using exampies, surface expressions, topics and loci. Unlike conventional work which was semantic markers for semantic constraints, we used examples for semantic constraints and showed in our experiments that examples are as useful as semantic markers. We also propose many new methods for estimating referents of'pronouns. For example, we use the form "X of Y" for estimating referents of demonstrative adjectives. In addition to our new methods, we used many conventional methods. As a result, experiments using these methods obtained a precision rate of 87% in estimating referents of demonstrative pronouns, personal pronouns, and zero pronouns for training sentences, and obtained a precision rate of 78% for test sentences. | Pronoun Resolution in Japanese Sentences Using Surface Expressions and Examples |
d1868 | We present an empirical study of the applicability of Probabilistic Lexicalized Tree Insertion Grammars (PLTIG), a lexicalized counterpart to Probabilistic Context-Free Grammars (PCFG), to problems in stochastic naturallanguage processing. Comparing the performance of PLTIGs with non-hierarchical N-gram models and PCFGs, we show that PLTIG combines the best aspects of both, with language modeling capability comparable to N-grams, and improved parsing performance over its nonlexicalized counterpart. Furthermore, training of PLTIGs displays faster convergence than PCFGs. | An Empirical Evaluation of Probabilistic Lexicalized Tree Insertion Grammars * |
d4061916 | Automatic interpretation of the relation between the constituents of a noun compound, e.g. olive oil (source) and baby oil (purpose) is an important task for many NLP applications. Recent approaches are typically based on either noun-compound representations or paraphrases. While the former has initially shown promising results, recent work suggests that the success stems from memorizing single prototypical words for each relation. We explore a neural paraphrasing approach that demonstrates superior performance when such memorization is not possible. | Olive Oil is Made of Olives, Baby Oil is Made for Babies: Interpreting Noun Compounds using Paraphrases in a Neural Model |
d146120598 | We investigate the recently developed Bidirectional Encoder Representations from Transformers (BERT) model(Devlin et al., 2018)for the hyperpartisan news detection task. Using a subset of hand-labeled articles from Se-mEval as a validation set, we test the performance of different parameters for BERT models. We find that accuracy from two different BERT models using different proportions of the articles is consistently high, with our bestperforming model on the validation set achieving 85% accuracy and the best-performing model on the test set achieving 77%. We further determined that our model exhibits strong consistency, labeling independent slices of the same article identically. Finally, we find that randomizing the order of word pieces dramatically reduces validation accuracy (to approximately 60%), but that shuffling groups of four or more word pieces maintains an accuracy of about 80%, indicating the model mainly gains value from local context. | Harvey Mudd College at SemEval-2019 Task 4: The Clint Buchanan Hyperpartisan News Detector |
d1540379 | We propose a language-independent method for the automatic extraction of transliteration pairs from parallel corpora. In contrast to previous work, our method uses no form of supervision, and does not require linguistically informed preprocessing. We conduct experiments on data sets from the NEWS 2010 shared task on transliteration mining and achieve an F-measure of up to 92%, outperforming most of the semi-supervised systems that were submitted. We also apply our method to English/Hindi and English/Arabic parallel corpora and compare the results with manually built gold standards which mark transliterated word pairs. Finally, we integrate the transliteration module into the GIZA++ word aligner and evaluate it on two word alignment tasks achieving improvements in both precision and recall measured against gold standard word alignments. | An Algorithm for Unsupervised Transliteration Mining with an Application to Word Alignment |
d10627917 | In this paper it is shown how simple texts that can be parsed in a Lambek Categorial Grammar can also automatically be provided with a semantics in the form of a Discourse Representation Structure in the sense ofKamp [1981]. The assignment of meanings to texts uses the Curry-Howard-Van Benthem correspondence. | CATEGORIAL GRAMMAR AND DISCOURSE REPRESENTATION THEORY |
d1550080 | Extracting summaries via integer linear programming and submodularity are popular and successful techniques in extractive multi-document summarization. However, many interesting optimization objectives are neither submodular nor factorizable into an integer linear program. We address this issue and present a general optimization framework where any function of input documents and a system summary can be plugged in. Our framework includes two kinds of summarizers -one based on genetic algorithms, the other using a swarm intelligence approach. In our experimental evaluation, we investigate the optimization of two information-theoretic summary evaluation metrics and find that our framework yields competitive results compared to several strong summarization baselines. Our comparative analysis of the genetic and swarm summarizers reveals interesting complementary properties. This work is licenced under a Creative Commons Attribution 4.0 International License. License details: http:// creativecommons.org/licenses/by/4.0/ | A General Optimization Framework for Multi-Document Summarization Using Genetic Algorithms and Swarm Intelligence |
d71028 | We show that in modeling social interaction, particularly dialogue, the attitude of obligation can be a useful adjunct to the popularly considered attitudes of belief, goal, and intention and their mutual and shared counterparts.In particular, we show how discourse obligations can be used to account in a natural manner for the connection between a question and its answer in dialogue and how obligations can be used along with other parts of the discourse context to extend the coverage of a dialogue system. | Discourse Obligations in Dialogue Processing |
d263609492 | We present our solution for the Russian RDFto-text generation task of the WebNLG Challenge 2023 1 .We use the pretrained large language model named FRED-T5(Zmitrovich et al., 2023)to finetune on the train dataset.Also, we propose several types of prompt and run experiments to analyze their effectiveness.Our submission achieves 0.373 TER on the test dataset, taking the first place according to the results of the automatic evaluation and outperforming the best result of the previous challenge by 0.025.The code of our solution is available at the following link: https: //github.com/Ivan30003/webnlg_interno | WebNLG-Interno: Utilizing FRED-T5 to address the RDF-to-text problem |
d870921 | We propose a new method of classifying documents into categories. We define for each category a finite mixture model based on soft clustering of words. We treat the problem of classifying documents as that of conducting statistical hypothesis testing over finite mixture models, and employ the EM algorithm to efficiently estimate parameters in a finite mixture model. Experimental results indicate that our method outperforms existing methods. | Document Classification Using a Finite Mixture Model |
d174800106 | A recently proposed lattice model has demonstrated that words in character sequence can provide rich word boundary information for character-based Chinese NER model. In this model, word information is integrated into a shortcut path between the start and the end characters of the word. However, the existence of shortcut path may cause the model to degenerate into a partial word-based model, which will suffer from word segmentation errors. Furthermore, the lattice model can not be trained in batches due to its DAG structure. In this paper, we propose a novel wordcharacter LSTM(WC-LSTM) model to add word information into the start or the end character of the word, alleviating the influence of word segmentation errors while obtaining the word boundary information. Four different strategies are explored in our model to encode word information into a fixed-sized representation for efficient batch training. Experiments on benchmark datasets show that our proposed model outperforms other stateof-the-arts models. | An Encoding Strategy Based Word-Character LSTM for Chinese NER |
d254854220 | Most research on task oriented dialog modeling is based on written text input. However, users interact with practical dialog systems often using speech as input. Typically, systems convert speech into text using an Automatic Speech Recognition (ASR) system, introducing errors. Furthermore, these systems do not address the differences in written and spoken language. The research on this topic is stymied by the lack of a public corpus. Motivated by these considerations, our goal in hosting the speech-aware dialog state tracking challenge was to create a public corpus or task which can be used to investigate the performance gap between the written and spoken forms of input, develop models that could alleviate this gap, and establish whether Textto-Speech-based (TTS) systems is a reasonable surrogate to the more-labor intensive human data collection. We created three spoken versions of the popular written-domain Mul-tiWoz task -(a) TTS-Verbatim: written user inputs were converted into speech waveforms using a TTS system, (b) Human-Verbatim: humans spoke the user inputs verbatim, and (c) Human-paraphrased: humans paraphrased the user inputs. Additionally, we provided different forms of ASR output to encourage wider participation from teams that may not have access to state-of-the-art ASR systems. These included ASR transcripts, word time stamps, and latent representations of the audio (audio encoder outputs). In this paper, we describe the corpus, report results from participating teams, provide preliminary analyses of their results, and summarize the current state-of-the-art in this domain. | Speech Aware Dialog System Technology Challenge (DSTC11) |
d5536079 | Semantic clusters of a domain form an important feature that can be useful for performing syntactic and semantic disambiguation. Several attempts have been made to extract the semantic clusters of a domain by probabilistic or taxonomic techniques. However, not much progress has been made in evaluating the obtained semantic clusters. This paper focuses on an evaluation mechanism that can be used to evaluate semantic clusters produced by a system against those provided by human experts. | Evaluation of Semantic Clusters |
d246823215 | Due to the high costs associated with finetuning large language models, various recent works propose to adapt them to specific tasks without any parameter updates through incontext learning. Unfortunately, for in-context learning there is currently no way to leverage unlabeled data, which is often much easier to obtain in large quantities than labeled examples. In this work, we therefore investigate ways to make use of unlabeled examples to improve the zero-shot performance of pretrained language models without any finetuning: We introduce Semantic-Oriented Unlabeled Priming (SOUP), a method that classifies examples by retrieving semantically similar unlabeled examples, assigning labels to them in a zero-shot fashion, and then using them for in-context learning. We also propose bag-of-contexts priming, a new priming strategy that is more suitable for our setting and enables the usage of more examples than fit into the context window. Sewon Min, Mike Lewis, Hannaneh Hajishirzi, and Luke Zettlemoyer. 2021. Noisy channel language model prompting for few-shot text classification. Computing Research Repository, arXiv:2108.04106. . 2020. Exploring the limits of transfer learning with a unified text-totext transformer. | Semantic-Oriented Unlabeled Priming for Large-Scale Language Models |
d9662312 | This paper describes a preliminary analysis of issues involved in the production of reports aimed at patients from Electronic Patient Records. We present a system prototype and discuss the problems encountered. | Exploring the Use of NLP in the Disclosure of Electronic Patient Records |
d53083475 | Legal Judgment Prediction (LJP) aims to predict the judgment result based on the facts of a case and becomes a promising application of artificial intelligence techniques in the legal field. In real-world scenarios, legal judgment usually consists of multiple subtasks, such as the decisions of applicable law articles, charges, fines, and the term of penalty. Moreover, there exist topological dependencies among these subtasks. While most existing works only focus on a specific subtask of judgment prediction and ignore the dependencies among subtasks, we formalize the dependencies among subtasks as a Directed Acyclic Graph (DAG) and propose a topological multi-task learning framework, TOP-JUDGE, which incorporates multiple subtasks and DAG dependencies into judgment prediction. We conduct experiments on several realworld large-scale datasets of criminal cases in the civil law system. Experimental results show that our model achieves consistent and significant improvements over baselines on all judgment prediction tasks. The source code can be obtained from https://github. com/thunlp/TopJudge. * Indicates equal contribution. The order is determined by dice rolling. † Corresponding author. On the early morning of July 24, 2017, the defendant XX stole cash 8500 yuan and T-shirts, jackets, pants, shoes, hats (identified a total value of 574.2 yuan) in Beijing Lining store… Law Article 264: [The crime of theft] Whoever steals a relatively large amount of public or private property or commits theft repeatedly fixed-term imprisonment of not more than three years, criminal detention or public surveillance. | Legal Judgment Prediction via Topological Learning |
d52157228 | We consider the cross-domain sentiment classification problem, where a sentiment classifier is to be learned from a source domain and to be generalized to a target domain. Our approach explicitly minimizes the distance between the source and the target instances in an embedded feature space. With the difference between source and target minimized, we then exploit additional information from the target domain by consolidating the idea of semi-supervised learning, for which, we jointly employ two regularizations -entropy minimization and self-ensemble bootstrapping -to incorporate the unlabeled target data for classifier refinement. Our experimental results demonstrate that the proposed approach can better leverage unlabeled data from the target domain and achieve substantial improvements over baseline methods in various experimental settings. | Adaptive Semi-supervised Learning for Cross-domain Sentiment Classification |
d221761373 | Adversarial attacks reveal important vulnerabilities and flaws of trained models. One potent type of attack are universal adversarial triggers, which are individual n-grams that, when appended to instances of a class under attack, can trick a model into predicting a target class. However, for inference tasks such as fact checking, these triggers often inadvertently invert the meaning of instances they are inserted in. In addition, such attacks produce semantically nonsensical inputs, as they simply concatenate triggers to existing samples. Here, we investigate how to generate adversarial attacks against fact checking systems that preserve the ground truth meaning and are semantically valid. We extend the HotFlip attack algorithm used for universal trigger generation by jointly minimizing the target class loss of a fact checking model and the entailment class loss of an auxiliary natural language inference model. We then train a conditional language model to generate semantically valid statements, which include the found universal triggers. We find that the generated attacks maintain the directionality and semantic validity of the claim better than previous work. . 2013. Intriguing properties of neural networks. arXiv preprint arXiv:1312.6199. | Generating Label Cohesive and Well-Formed Adversarial Claims |
d258762836 | State-of-the-art target-oriented opinion word extraction (TOWE) models typically use BERTbased text encoders that operate on the word level, along with graph convolutional networks (GCNs) that incorporate syntactic information extracted from syntax trees. These methods achieve limited gains with GCNs and have difficulty using BERT wordpieces. Meanwhile, BERT wordpieces are known to be effective at representing rare words or words with insufficient context information. To address this issue, this work trades syntax trees for BERT wordpieces by entirely removing the GCN component from the methods' architectures. To enhance TOWE performance, we tackle the issue of aspect representation loss during encoding. Instead of solely utilizing a sentence as the input, we use a sentence-aspect pair. Our relatively simple approach achieves state-of-the-art results on benchmark datasets and should serve as a strong baseline for further research. | Trading Syntax Trees for Wordpieces: Target-oriented Opinion Words Extraction with Wordpieces and Aspect Enhancement |
d260680487 | In this work, we study dialogue scenarios that start from chit-chat but eventually switch to task-related services, and investigate how a unified dialogue model, which can engage in both chit-chat and task-oriented dialogues, takes the initiative during the dialogue mode transition from chit-chat to task-oriented in a coherent and cooperative manner.We firstly build a transition info extractor (TIE) that keeps track of the preceding chit-chat interaction and detects the potential user intention to switch to a taskoriented service.Meanwhile, in the unified model, a transition sentence generator (TSG) is extended through efficient Adapter tuning and transition prompt learning.When the TIE successfully finds task-related information from the preceding chit-chat, such as a transition domain ("train" in Figure1), then the TSG is activated automatically in the unified model to initiate this transition by generating a transition sentence under the guidance of transition information extracted by TIE.The experimental results show promising performance regarding the proactive transitions.We achieve an additional large improvement on TIE model by utilizing Conditional Random Fields (CRF).The TSG can flexibly generate transition sentences while maintaining the unified capabilities of normal chit-chat and task-oriented response generation. | System-Initiated Transitions from Chit-Chat to Task-Oriented Dialogues with Transition Info Extractor and Transition Sentence Generator |
d1459907 | Named entity recognition, which focuses on the identification of the span and type of named entity mentions in texts, has drawn the attention of the NLP community for a long time. However, many real-life applications need to know which real entity each mention refers to. For such a purpose, often refered to as entity resolution and linking, an inventory of entities is required in order to constitute a reference. In this paper, we describe how we extracted such a resource for French from freely available resources (the French Wikipedia and the GeoNames database). We describe the results of an instrinsic evaluation of the resulting entity database, named Aleda, as well as those of a task-based evaluation in the context of a named entity detection system. We also compare it with the NLGbAse database (Charton and Torres-Moreno, 2010), a resource with similar objectives. | Aleda, a free large-scale entity database for French |
d241033035 | This report describes Microsoft's machine translation systems for the WMT21 shared task on large-scale multilingual machine translation. We participated in all three evaluation tracks including Large Track and two Small Tracks where the former one is unconstrained and the latter two are fully constrained. Our model submissions to the shared task were initialized with DeltaLM 1 , a generic pre-trained multilingual encoder-decoder model, and finetuned correspondingly with the vast collected parallel data and allowed data sources according to track settings, together with applying progressive learning and iterative backtranslation approaches to further improve the performance. Our final submissions ranked first on three tracks in terms of the automatic evaluation metric. | Multilingual Machine Translation Systems from Microsoft for WMT21 Shared Task |
d263883798 | We describe Mega-COV, a billion-scale dataset from Twitter for studying COVID-19. The dataset is diverse (covers 268 countries), longitudinal (goes as back as 2007), multilingual (comes in 100+ languages), and has a significant number of location-tagged tweets (∼ 169M tweets). We release tweet IDs from the dataset. We also develop two powerful models, one for identifying whether or not a tweet is related to the pandemic (best F 1 =97%) and another for detecting misinformation about COVID-19 (best F 1 =92%). A human annotation study reveals the utility of our models on a subset of Mega-COV. Our data and models can be useful for studying a wide host of phenomena related to the pandemic. Mega-COV and our models are publicly available.Tomer Simon, Avishay Goldberg, and Bruria Adini.2015. Socializing in emergencies-a review of the use of social media in emergency situations. | Mega-COV: A Billion-Scale Dataset of 100+ Languages for COVID-19 |
d14839256 | Existing algorithms for the Generation of Referring Expressions tend to generate distinguishing descriptions at the semantic level, disregarding the ways in which surface issues can affect their quality. This paper considers how these algorithms should deal with surface ambiguity, focussing on structural ambiguity. We propose that not all ambiguity is worth avoiding, and suggest some ways forward that attempt to avoid unwanted interpretations. We sketch the design of an algorithm motivated by our experimental findings. | Generation of Referring Expressions: Managing Structural Ambiguities * |
d337425 | We present DEPCC, the largest-to-date linguistically analyzed corpus in English including 365 million documents, composed of 252 billion tokens and 7.5 billion of named entity occurrences in 14.3 billion sentences from a web-scale crawl of the COMMON CRAWL project. The sentences are processed with a dependency parser and with a named entity tagger and contain provenance information, enabling various applications ranging from training syntax-based word embeddings to open information extraction and question answering. We built an index of all sentences and their linguistic meta-data enabling quick search across the corpus. We demonstrate the utility of this corpus on the verb similarity task by showing that a distributional model trained on our corpus yields better results than models trained on smaller corpora, like Wikipedia. This distributional model outperforms the state of art models of verb similarity trained on smaller corpora on the SimVerb3500 dataset. | Building a Web-Scale Dependency-Parsed Corpus from Common Crawl |
d6462501 | This paper describes the AMU-UEDIN submissions to the WMT 2016 shared task on news translation. We explore methods of decode-time integration of attention-based neural translation models with phrase-based statistical machine translation. Efficient batch-algorithms for GPU-querying are proposed and implemented. For English-Russian, our system stays behind the state-of-the-art pure neural models in terms of BLEU. Among restricted systems, manual evaluation places it in the first cluster tied with the pure neural model. For the Russian-English task, our submission achieves the top BLEU result, outperforming the best pure neural system by 1.1 BLEU points and our own phrase-based baseline by 1.6 BLEU. After manual evaluation, this system is the best restricted system in its own cluster. In follow-up experiments we improve results by additional 0.8 BLEU. | Shared Task Papers |
d220444944 | For mining intellectual property texts (patents), a broad-coverage lexicon that covers general English words together with terminology from the patent domain is indispensable. The patent domain is very diffuse as it comprises a variety of technical domains (e.g. Human Necessities, Chemistry & Metallurgy and Physics in the International Patent Classification). As a result, collecting a lexicon that covers the language used in patent texts is not a straightforward task. In this paper we describe the approach that we have developed for the semi-automatic construction of a broad-coverage lexicon for classification and information retrieval in the patent domain and which combines information from multiple sources. Our contribution is twofold. First, we provide insight into the difficulties of developing lexical resources for information retrieval and text mining in the patent domain, a research and development field that is expanding quickly. Second, we create a broad coverage lexicon annotated with rich lexical information and containing both general English word forms and domain terminology for various technical domains. | Constructing a broad-coverage lexicon for text mining in the patent domain |
d52874710 | We present a language-independent and unsupervised algorithm for the segmentation of words into morphs. The algorithm is based on a new generative probabilistic model, which makes use of relevant prior information on the length and frequency distributions of morphs in a language. Our algorithm is shown to outperform two competing algorithms, when evaluated on data from a language with agglutinative morphology (Finnish), and to perform well also on English data. | Unsupervised Segmentation of Words Using Prior Distributions of Morph Length and Frequency |
d225062645 | As a core task of Information Extraction, Entity Relation Extraction plays an important role in many Natural Language Processing applications such as knowledge graph, intelligent question answering system and semantic search. Relation extraction tasks aim to find the semantic relation between a pair of entity mentions from unstructured texts. This paper focuses on the sentence-level relation extraction, introduces the main datasets for this task, and expounds the current status of relation extraction technology which can be divided into: supervised relation extraction, distant supervision relation extraction and joint extraction of entities and relations. We compare the various models for this task and analyze their contributions and defects. Finally, the research status and methods of Chinese entity relation extraction are introduced. | Review of Entity Relation Extraction based on deep learning |
d8214692 | Computational models for sarcasm detection have often relied on the content of utterances in isolation. However, speaker's sarcastic intent is not always obvious without additional context. Focusing on social media discussions, we investigate two issues: (1) does modeling of conversation context help in sarcasm detection and(2)can we understand what part of conversation context triggered the sarcastic reply. To address the first issue, we investigate several types of Long Short-Term Memory (LSTM) networks that can model both the conversation context and the sarcastic response. 1 We show that the conditional LSTM network(Rocktäschel et al., 2015)and LSTM networks with sentence level attention on context and response outperform the LSTM model that reads only the response. To address the second issue, we present a qualitative analysis of attention weights produced by the LSTM models with attention and discuss the results compared with human performance on the task. | The Role of Conversation Context for Sarcasm Detection in Online Interactions |
d209370514 | Current approaches to machine translation (MT) either translate sentences in isolation, disregarding the context they appear in, or model context at the level of the full document, without a notion of any internal structure the document may have. In this work we consider the fact that documents are rarely homogeneous blocks of text, but rather consist of parts covering different topics. Some documents, such as biographies and encyclopedia entries, have highly predictable, regular structures in which sections are characterised by different topics. We draw inspiration from Louis and Webber (2014) who use this information to improve statistical MT and transfer their proposal into the framework of neural MT. We compare two different methods of including information about the topic of the section within which each sentence is found: one using side constraints and the other using a cache-based model. We create and release the data on which we run our experiments -parallel corpora for three language pairs (Chinese-English, French-English, Bulgarian-English) from Wikipedia biographies, which we extract automatically, preserving the boundaries of sections within the articles. | Document Sub-structure in Neural Machine Translation |
d52114113 | We release a corpus of 43 million atomic edits across 8 languages. These edits are mined from Wikipedia edit history and consist of instances in which a human editor has inserted a single contiguous phrase into, or deleted a single contiguous phrase from, an existing sentence. We use the collected data to show that the language generated during editing differs from the language that we observe in standard corpora, and that models trained on edits encode different aspects of semantics and discourse than models trained on raw, unstructured text. We release the full corpus as a resource to aid ongoing research in semantics, discourse, and representation learning. | WikiAtomicEdits: A Multilingual Corpus of Wikipedia Edits for Modeling Language and Discourse |
d261705995 | Developing high-performing dialogue systems benefits from the automatic identification of undesirable behaviors in system responses.However, detecting such behaviors remains challenging, as it draws on a breadth of general knowledge and understanding of conversational practices.Although recent research has focused on building specialized classifiers for detecting specific dialogue behaviors, the behavior coverage is still incomplete and there is a lack of testing on real-world human-bot interactions.This paper investigates the ability of a state-of-the-art large language model (LLM), ChatGPT-3.5, to perform dialogue behavior detection for nine categories in real human-bot dialogues.We aim to assess whether ChatGPT can match specialized models and approximate human performance, thereby reducing the cost of behavior detection tasks.Our findings reveal that neither specialized models nor Chat-GPT have yet achieved satisfactory results for this task, falling short of human performance.Nevertheless, ChatGPT shows promising potential and often outperforms specialized detection models.We conclude with an in-depth examination of the prevalent shortcomings of ChatGPT, offering guidance for future research to enhance LLM capabilities. | Leveraging Large Language Models for Automated Dialogue Analysis |
d1998866 | This paper presents the approach of the GTI Research Group to SemEval-2015 task 10 on Sentiment Analysis in Twitter, or more specifically, subtasks A (Contextual Polarity Disambiguation) and B (Message Polarity Classification). We followed an unsupervised dependency parsing-based approach using a sentiment lexicon, created by means of an automatic polarity expansion algorithm and Natural Language Processing techniques. These techniques involve the use of linguistic peculiarities, such as the detection of polarity conflicts or adversative/concessive subordinate clauses. The results obtained confirm the competitive and robust performance of the system. | GTI: An Unsupervised Approach for Sentiment Analysis in Twitter |
d17865105 | Automatically generating product reviews is a meaningful, yet not well-studied task in sentiment analysis. Traditional natural language generation methods rely extensively on hand-crafted rules and predefined templates. This paper presents an attention-enhanced attribute-to-sequence model to generate product reviews for given attribute information, such as user, product, and rating. The attribute encoder learns to represent input attributes as vectors. Then, the sequence decoder generates reviews by conditioning its output on these vectors. We also introduce an attention mechanism to jointly generate reviews and align words with input attributes. The proposed model is trained end-to-end to maximize the likelihood of target product reviews given the attributes. We build a publicly available dataset for the review generation task by leveraging the Amazon book reviews and their metadata. Experiments on the dataset show that our approach outperforms baseline methods and the attention mechanism significantly improves the performance of our model. | Learning to Generate Product Reviews from Attributes |
d209387655 | Surface realisation maps a meaning representation (MR) to a text, usually a single sentence. In this paper, we introduce a new parallel dataset of deep meaning representations and French sentences and we present a novel method for MR-to-text generation which seeks to generalise by abstracting away from lexical content. Most current work on natural language generation focuses on generating text that matches a reference using BLEU as evaluation criteria. In this paper, we additionally consider the model's ability to reintroduce the function words that are absent from the deep input meaning representations. We show that our approach increases both BLEU score and the scores used to assess function words generation. | Generating Text from Anonymised Structures |
d4708673 | In this paper, we present a kernel-based learning approach for the 2018 Complex Word Identification (CWI) Shared Task. Our approach is based on combining multiple lowlevel features, such as character n-grams, with high-level semantic features that are either automatically learned using word embeddings or extracted from a lexical knowledge base, namely WordNet. After feature extraction, we employ a kernel method for the learning phase. The feature matrix is first transformed into a normalized kernel matrix. For the binary classification task (simple versus complex), we employ Support Vector Machines. For the regression task, in which we have to predict the complexity level of a word (a word is more complex if it is labeled as complex by more annotators), we employ ν-Support Vector Regression. We applied our approach only on the three English | UnibucKernel: A kernel-based learning method for complex word identification |
d253107178 | Question answering models can use rich knowledge sources -up to one hundred retrieved passages and parametric knowledge in the large-scale language model (LM). Prior work assumes information in such knowledge sources is consistent with each other, paying little attention to how models blend information stored in their LM parameters with that from retrieved evidence documents. In this paper, we simulate knowledge conflicts (i.e., where parametric knowledge suggests one answer and different passages suggest different answers) and examine model behaviors. We find retrieval performance heavily impacts which sources models rely on, and current models mostly rely on non-parametric knowledge in their best-performing settings. We discover a troubling trend that contradictions among knowledge sources affect model confidence only marginally. To address this issue, we present a new calibration study, where models are discouraged from presenting any single answer when presented with multiple conflicting answer candidates in retrieved evidences. | Rich Knowledge Sources Bring Complex Knowledge Conflicts: Recalibrating Models to Reflect Conflicting Evidence |
d2083283 | Spoken dialogue systems promise efficient and natural access to information services from any phone. Recently, spoken dialogue systems for widely used applications such as email, travel information, and customer care have moved from research labs into commercial use. These applications can receive millions of calls a month. This huge amount of spoken dialogue data has led to a need for fully automatic methods for selecting a subset of caller dialogues that are most likely to be useful for further system improvement, to be stored, transcribed and further analyzed. This paper reports results on automatically training a Problematic Dialogue Identifier to classify problematic human-computer dialogues using a corpus of 1242 DARPA Communicator dialogues in the travel planning domain. We show that using fully automatic features we can identify classes of problematic dialogues with accuracies from 67% to 89%. | What's the Trouble: Automatically Identifying Problematic Dialogues in DARPA Communicator Dialogue Systems |
d3032231 | In this paper we present the first step in a larger series of experiments for the induction of predicate/argument structures. The structures that we are inducing are very similar to the conceptual structures that are used in Frame Semantics (such as FrameNet). Those structures are called messages and they were previously used in the context of a multi-document summarization system of evolving events. The series of experiments that we are proposing are essentially composed from two stages. In the first stage we are trying to extract a representative vocabulary of words. This vocabulary is later used in the second stage, during which we apply to it various clustering approaches in order to identify the clusters of predicates and arguments-or frames and semantic roles, to use the jargon of Frame Semantics. This paper presents in detail and evaluates the first stage. | What's in a Message? |
d4659219 | Concept maps can be used to concisely represent important information and bring structure into large document collections. Therefore, we study a variant of multidocument summarization that produces summaries in the form of concept maps. However, suitable evaluation datasets for this task are currently missing. To close this gap, we present a newly created corpus of concept maps that summarize heterogeneous collections of web documents on educational topics. It was created using a novel crowdsourcing approach that allows us to efficiently determine important elements in large document collections. We release the corpus along with a baseline system and proposed evaluation protocol to enable further research on this variant of summarization. 1 | Bringing Structure into Summaries: Crowdsourcing a Benchmark Corpus of Concept Maps |
d245855869 | This paper describes our submission to the WMT2021 shared metrics task. Our metric is operative to segment-level and system-level translations. Our belief toward a better metric is to detect a significant error that cannot be missed in the real practice cases of evaluation. For that reason, we used pseudo-negative examples in which attributes of some words are transferred to the reversed attribute words, and we build evaluation models to handle such serious mistakes of translations. We fine-tune a multilingual largely pre-trained model on the provided corpus of past years' metric task and fine-tune again further on the synthetic negative examples that are derived from the same fine-tune corpus. From the evaluation results of the WMT21's development corpus, finetuning on the pseudo-negatives using WMT15-17 and WMT18-20 metric corpus achieved a better Pearson's correlation score than the one fine-tuned without negative examples. Our submitted models are named C-SPEC (Crosslingual Sentence Pair Embedding Concatenation) and C-SPECpn, are the plain model using WMT18-20 and the one additionally finetuned on negative samples, respectively. | Multilingual Machine Translation Evaluation Metrics Fine-tuned on Pseudo-Negative Examples for WMT 2021 Metrics Task |
d2142405 | We present two related tasks of the BioNLP Shared Tasks 2011: Bacteria Gene Renaming (Rename) and Bacteria Gene Interactions (GI). We detail the objectives, the corpus specification, the evaluation metrics, and we summarize the participants' results. Both issued from PubMed scientific literature abstracts, the Rename task aims at extracting gene name synonyms, and the GI task aims at extracting genic interaction events, mainly about gene transcriptional regulations in bacteria. | BioNLP Shared Task 2011 -Bacteria Gene Interactions and Renaming |
d259833805 | It is well known that filtering low-quality data before pretraining language models or selecting suitable data from domains similar to downstream task datasets generally leads to improved downstream performance. However, the extent to which the quality of a corpus, in particular its complexity, affects its downstream performance remains less explored. In this work, we address the problem of creating a suitable pretraining corpus given a fixed corpus budget. Using metrics of text complexity we propose a simple yet effective approach for constructing a corpus with rich lexical variation. Our extensive set of empirical analyses reveal that such a diverse and complex corpus yields significant improvements over baselines consisting of less diverse and less complex corpora when evaluated in the context of general language understanding tasks. | Corpus Complexity Matters in Pretraining Language Models |
d193020094 | RÉSUMÉ________________________________________________________________________Cet article aborde la problématique de l'annotation automatique d'un corpus d'apprenants d'anglais. L'objectif est de montrer qu'il est possible d'utiliser un étiqueteur PoS pour annoter un corpus d'apprenants afin d'analyser les erreurs faites par les apprenants. Cependant, pour permettre une analyse suffisamment fine, des étiquettes fonctionnelles spécifiques aux phénomènes linguistiques à étudier sont insérées parmi celles de l'étiqueteur. Celui-ci est entraîné avec ce jeu d'étiquettes étendu sur un corpus de natifs avant d'être appliqué sur le corpus d'apprenants. Dans cette expérience, on s'intéresse aux usages erronés de this et that par les apprenants. On montre comment l'ajout d'une couche fonctionnelle sous forme de nouvelles étiquettes pour ces deux formes, permet de discriminer des usages variables chez les natifs et non-natifs et, partant, d'identifier des schémas incorrects d'utilisation. Les étiquettes fonctionnelles éclairent sur le fonctionnement discursif.ABSTRACT_______________________________________________________________________Automatic tagging of a learner corpus of English with a modified version of the Penn Treebank tagsetThis article covers the issue of automatic annotation of a learner corpus of English. The objective is to show that it is possible to PoS-tag the corpus with a tagger to prepare the ground for learner error analysis. However, in order to have a fine-grain analysis, some functional tags for the study of specific linguistic points are inserted within the tagger's tagset. This tagger is trained on a native-English corpus with an extended tagset and the tagging is done on the learner corpus. This experiment focuses on the incorrect use of this and that by learners. We show how the insertion of a functional layer by way of new tags for the forms allows us to discriminate varying uses among natives and non-natives. This opens the path to the identification of incorrect patterns of use. The functional tags cast a light on the way the discourse functions.MOTS-CLÉS : Apprentissage L2, corpus d'apprenants, analyse linguistique d'erreurs, étiquetage automatique, this, that | Annotation automatique d'un corpus d'apprenants d'anglais avec un jeu d'étiquettes modifié du Penn Treebank |
d184483883 | We present the Named Entity Recognition (NER) and disambiguation model used by the University of Arizona team (UArizona) for SemEval 2019 task 12. We achieved fourth place on tasks 1 and 3. We implemented a deep-affix based LSTM-CRF NER model for task 1, which utilizes only character, word, prefix and suffix information for the identification of geolocation entities. Despite using just the training data provided by task organizers and not using any lexicon features, we achieved 78.85% strict micro F-score on task 1. We used the unsupervised population heuristics for task 3 and achieved 52.99% strict micro-F1 score in this task. | Deep-Affix Named Entity Recognition of Geolocation Entities |
d16108530 | This paper presents a novel method for aligning etymological data, which models context-sensitive rules governing sound change, and utilizes phonetic features of the sounds. The goal is, for a given corpus of cognate sets, to find the best alignment at the sound level. We introduce an imputation procedure to compare the goodness of the resulting models, as well as the goodness of the data sets. We present evaluations to demonstrate that the new model yields improvements in performance, compared to previously reported models. | Using context and phonetic features in models of etymological sound change |
d3174922 | We present an extension of the adverbial entries of the French morphological lexicon DELA (Dictionnaires Electroniques du LADL / LADL electronic dictionaries). Adverbs were extracted from LGLex, a NLP-oriented syntactic resource for French, which in its turn contains all adverbs extracted from the Lexicon-Grammar tables of both simple adverbs ending in -ment (i.e., '-ly')(Molinier and Levrier, 2000)and compound adverbs(Gross, 1986b;Gross, 1986a). This work exploits fine-grained linguistic information provided in existing resources. The resulting resource is reviewed in order to delete duplicates and is freely available under the LGPL-LR license. | Extending the adverbial coverage of a French morphological lexicon |
d49208337 | Mental health is a significant and growing public health concern. As language usage can be leveraged to obtain crucial insights into mental health conditions, there is a need for large-scale, labeled, mental health-related datasets of users who have been diagnosed with one or more of such conditions. In this paper, we investigate the creation of high-precision patterns to identify self-reported diagnoses of nine different mental health conditions, and obtain high-quality labeled data without the need for manual labelling. We introduce the SMHD (Self-reported Mental Health Diagnoses) dataset and make it available. SMHD is a novel large dataset of social media posts from users with one or multiple mental health conditions along with matched control users. We examine distinctions in users' language, as measured by linguistic and psychological variables. We further explore text classification methods to identify individuals with mental conditions through their language. * Equal contribution. This work is licensed under a Creative Commons Attribution 4.0 International License. License details: http:// creativecommons.org/licenses/by/4.0/ | SMHD: A Large-Scale Resource for Exploring Online Language Usage for Multiple Mental Health Conditions |
d1726328 | This paper describes the Duluth systems that participated in SemEval-2017 Task 7 : Detection and Interpretation of English Puns. The Duluth systems participated in all three subtasks, and relied on methods that included word sense disambiguation and measures of semantic relatedness. | Duluth at SemEval-2017 Task 7: Puns Upon a Midnight Dreary, Lexical Semantics for the Weak and Weary |
d498 | Excellent results have been reported for Data-Oriented Parsing (DOP) of natural language texts(Bod, 1993c). Unfortunately, existing algorithms are both computationally intensive and difficult to implement. Previous algorithms are expensive due to two factors: the exponential number of rules that must be generated and the use of a Monte Carlo parsing algorithm. In this paper we solve the first problem by a novel reduction of the DOP model to:a small, equivalent probabilistic context-free grammar. We solve the second problem by a novel deterministic parsing strategy that maximizes the expected number of correct constituents, rather than the probability of a correct parse tree. Using ithe optimizations, experiments yield a 97% crossing brackets rate and 88% zero crossing brackets rate. This differs significantly from the results reported by Bod, and is comparable to results from a duplication of Pereira and Schabes's (1992) experiment on the same data. We show that Bod's results are at least partially due to an extremely fortuitous choice of test data, and partially due to using cleaner data than other researchers. | Efficient Algorithms for Parsing the DOP Model * |
d3264822 | This paper presents a method for the automatic extraction of subgrammars to control and speeding-up natural language generation NLG. The method is based on explanation-based learning EBL. The main advantage for the proposed new method for NLG is that the complexity of the grammatical decision making process during NLG can be vastly reduced, because the EBL method supports the adaption of a NLG system to a particular use of a language. | Applying Explanation-based Learning to Control and Speeding-up Natural Language Generation |
d237490383 | Creole languages such as Nigerian Pidgin English and Haitian Creole are under-resourced and largely ignored in the NLP literature. Creoles typically result from the fusion of a foreign language with multiple local languages, and what grammatical and lexical features are transferred to the creole is a complex process (Sessarego, 2020). While creoles are generally stable, the prominence of some features may be much stronger with certain demographics or in some linguistic situations(Winford, 1999;Patrick, 1999). This paper makes several contributions: We collect existing corpora and release models for Haitian Creole, Nigerian Pidgin English, and Singaporean Colloquial English. We evaluate these models on intrinsic and extrinsic tasks. Motivated by the above literature, we compare standard language models with distributionally robust ones and find that, somewhat surprisingly, the standard language models are superior to the distributionally robust ones. We investigate whether this is an effect of overparameterization or relative distributional stability, and find that the difference persists in the absence of over-parameterization, and that drift is limited, confirming the relative stability of creole languages. | On Language Models for Creoles |
d252519231 | Multiple business scenarios require an automated generation of descriptive human-readable text from structured input data. Hence, fact-to-text generation systems have been developed for various downstream tasks like generating soccer reports, weather and financial reports, medical reports, person biographies, etc. Unfortunately, previous work on fact-to-text (F2T) generation has focused primarily on English mainly due to the high availability of relevant datasets. Only recently, the problem of cross-lingual fact-to-text (XF2T) was proposed for generation across multiple languages alongwith a dataset, XALIGN for eight languages. However, there has been no rigorous work on the actual XF2T generation problem. We extend XALIGN dataset with annotated data for four more languages: Punjabi, Malayalam, Assamese and Oriya. We conduct an extensive study using popular Transformer-based text generation models on our extended multi-lingual dataset, which we call XALIGNV2. Further, we investigate the performance of different text generation strategies: multiple variations of pretraining, fact-aware embeddings and structure-aware input encoding. Our extensive experiments show that a multi-lingual mT5 model which uses fact-aware embeddings with structure-aware input encoding leads to best results on average across the twelve languages. We make our code, dataset and model publicly available 1 , and hope that this will help advance further research in this critical area. arXiv:2209.11252v1 [cs.CL] 22 Sep 2022 XF2T <Elon_Musk, nationality, South_Africa> <Elon_Musk, nationality, Canada > <Elon_Musk, nationality, USA> <Elon_Musk, date_of_birth, 28_June_1971 > <Elon_Musk, occupation, engineer> <Elon_Musk, occupation, entrepreneur> <Elon_Musk, occupation, inventor> <Elon_Musk, occupation, investor> <hindi> एलन म (ज 28 जू न 1971) एक दि ण अ ीकी-कनाडाई-अमे रकी िद ज ापारी, िनवे शक, इं जीिनयर, और आिव ारक ह । <bengali> এলন মা (জ 28 জু ন 1971) দি ণ আি কা-কানািডয়ান-আেমিরকান বীণ ব বসায়ী, িবিনেয়াগকারী, েকৗশলী এবং উ াবক। <tamil> எேலான் மஸ ் க் ( றப் 28 ஜ ன் 1971) ஒ ெதன் னாப் ரிக் க-கன ய-அெமரிக் க த் த ெதா ல பர் , த ட் டாளர் , ெபா யாளர் மற் ம் கண ் ப் பாளர் ஆவார் . English Facts <gujarati> એલોન મ ક (જ મ 28 જૂ ન 1971) એ દિ ણ આિ કા-કે ને િડયન-અમે િરકન પીte ઉ ોગપિત, રોકાણકાર, ઇજને ર અને શોધક છે . <English> Elon Musk (born 28 June 1971) is a South African-Canadian-American veteran businessman, investor, engineer, and inventor. <punjabi> ਐਲੋ ਨ ਮਸਕ (ਜਨਮ 28 ਜੂ ਨ 1971) ਇੱ ਕ ਦੱ ਖਣੀ ਅਫ਼ਰੀਕੀ-ਕੈ ਨੇ ਡੀਅਨ-ਅਮਰੀਕੀ ਅਨੁ ਭਵੀ ਕਾਰੋ ਬਾਰੀ, ਿਨਵੇ ਸ਼ਕ, ਇੰ ਜੀਨੀਅਰ, ਅਤੇ ਖੋ ਜੀ ਹੈ । ... | XF2T: Cross-lingual Fact-to-Text Generation for Low-Resource Languages |
d252547725 | Pretrained multilingual language models can help bridge the digital language divide, enabling high-quality NLP models for lowerresourced languages. Studies of multilingual models have so far focused on performance, consistency, and cross-lingual generalisation. However, with their wide-spread application in the wild and downstream societal impact, it is important to put multilingual models under the same scrutiny as monolingual models. This work investigates the group fairness of multilingual models, asking whether these models are equally fair across languages. To this end, we create a new four-way multilingual dataset of parallel cloze test examples (MozArt), equipped with demographic information (balanced with regard to gender and native tongue) about the test participants. We evaluate three multilingual models on MozArt -mBERT, XLM-R, and mT5 -and show that across the four target languages, the three models exhibit different levels of group disparity, e.g., exhibiting near-equal risk for Spanish, but high levels of disparity for German.John Schmitz. 2016. On the native/nonnative speaker notion and world englishes: Debating with k. rajagopalan. DELTA: Documentação de Estudos em Lingüística Teórica e Aplicada, 32:597-611. | Are Pretrained Multilingual Models Equally Fair Across Languages? |
d256460904 | Causal (a cause-effect relationship between two arguments) has become integral to various NLP domains such as question answering, summarization, and event prediction. To understand causality in detail, Event Causality Identification with Causal News Corpus (CASE-2022) has organized shared tasks. This paper defines our participation in Subtask 1, which focuses on classifying event causality. We used sentence level augmentation based on contextualized word embeddings of distillBERT to construct new data. This data was then trained using two approaches. The first technique used the DeBERTa language model, and the second used the RoBERTa language model in combination with cross attention. We obtained the second-best F1 score (0.8610) in the competition with Contextually Augmented DeBERTa model. | ARGUABLY @ Causal News Corpus 2022: Contextually Augmented Language Models for Event Causality Identification |
d3677429 | We introduce ParlAI (pronounced "par-lay"), an open-source software platform for dialog research implemented in Python, available at http://parl.ai. Its goal is to provide a unified framework for sharing, training and testing dialog models; integration of Amazon Mechanical Turk for data collection, human evaluation, and online/reinforcement learning; and a repository of machine learning models for comparing with others' models, and improving upon existing architectures. Over 20 tasks are supported in the first release, including popular datasets such as SQuAD, bAbI tasks, MCTest, WikiQA, QACNN, QADaily-Mail, CBT, bAbI Dialog, Ubuntu, OpenSubtitles and VQA. Several models are integrated, including neural models such as memory networks, seq2seq and attentive LSTMs. | ParlAI: A Dialog Research Software Platform |
d259370630 | Pre-trained autoregressive (AR) language models such as BART and GPTs have dominated Open-ended Long Text Generation (Open-LTG). However, the AR nature will decrease the inference efficiency along with the increase of generation length, which hinder their application in Open-LTG. To improve inference efficiency, we alternatively explore the potential of the pre-trained masked language models (MLMs) along with a representative iterative non-autoregressive (NAR) decoding strategy for Open-LTG. Our preliminary study shows that pre-trained MLMs can merely generate short text and will collapse for long text modeling. To enhance the long text generation capability of MLMs, we introduce two simple yet effective strategies for the iterative NAR model: dynamic sliding window attention (DSWA) and linear temperature decay (LTD). It can alleviate long-distance collapse problems and achieve longer text generation with a flexible trade-off between performance and inference speedup. Experiments on the storytelling and multi-paragraph opinionated article writing tasks show that pre-trained MLMs can achieve more than 3 × → 13 × speedup with better performance than strong AR models. Our code is available at GitHub * . . 2022. A survey on non-autoregressive generation for neural machine translation and beyond. arXiv preprint arXiv:2204.09269. | Open-ended Long Text Generation via Masked Language Modeling |
d254854405 | Identifying named entities such as a person, location or organization, in documents can highlight key information to readers. Training Named Entity Recognition (NER) models requires an annotated data set, which can be a time-consuming labour-intensive task. Nevertheless, there are publicly available NER data sets for general English. Recently there has been interest in developing NER for legal text. However, prior work and experimental results reported here indicate that there is a significant degradation in performance when NER methods trained on a general English data set are applied to legal text. We describe a publicly available legal NER data set, called E-NER, based on legal company filings available from the US Securities and Exchange Commission's EDGAR data set. Training a number of different NER algorithms on the general English CoNLL-2003 corpus but testing on our test collection confirmed significant degradations in accuracy, as measured by the F1-score, of between 29.4% and 60.4%, compared to training and testing on the E-NER collection. | E-NER -An Annotated Named Entity Recognition Corpus of Legal Text |
d53603184 | This paper describes the systems developed by IRISA to participate to the four tasks of the SMM4H 2018 challenge. For these tweet classification tasks, we adopt a common approach based on recurrent neural networks (BiLSTM). Our main contributions are the use of certain features, the use of Bagging in order to deal with unbalanced datasets, and on the automatic selection of difficult examples. These techniques allow us to reach 91.4, 46.5, 47.8, 85.0 as F1-scores for Tasks 1 to 4. | IRISA at SMM4H 2018: Neural Network and Bagging for Tweet Classification |
d263868453 | The canonical word order of Japanese double object constructions has attracted considerable attention among linguists and has been a topic of many studies. However, most of these studies require either manual analyses or measurements of human characteristics such as brain activities or reading times for each example. Thus, while these analyses are reliable for the examples they focus on, they cannot be generalized to other examples. On the other hand, the trend of actual usage can be collected automatically from a large corpus. Thus, in this paper, we assume that there is a relationship between the canonical word order and the proportion of each word order in a large corpus and present a corpusbased analysis of canonical word order of Japanese double object constructions. | A Corpus-Based Analysis of Canonical Word Order of Japanese Double Object Constructions |
d19021652 | We present a framework for the acquisition of sentential paraphrases based on crowdsourcing. The proposed method maximizes the lexical divergence between an original sentence s and its valid paraphrases by running a sequence of paraphrasing jobs carried out by a crowd of non-expert workers. Instead of collecting direct paraphrases of s, at each step of the sequence workers manipulate semantically equivalent reformulations produced in the previous round. We applied this method to paraphrase English sentences extracted from Wikipedia. Our results show that, keeping at each round n the most promising paraphrases (i.e. the more lexically dissimilar from those acquired at round n-1), the monotonic increase of divergence allows to collect good-quality paraphrases in a cost-effective manner. | Chinese Whispers: Cooperative Paraphrase Acquisition |
d8951658 | This paper proposes KB-InfoBot 1 -a multi-turn dialogue agent which helps users search Knowledge Bases (KBs) without composing complicated queries. Such goal-oriented dialogue agents typically need to interact with an external database to access real-world knowledge. Previous systems achieved this by issuing a symbolic query to the KB to retrieve entries based on their attributes. However, such symbolic operations break the differentiability of the system and prevent endto-end training of neural dialogue agents. In this paper, we address this limitation by replacing symbolic queries with an induced "soft" posterior distribution over the KB that indicates which entities the user is interested in. Integrating the soft retrieval process with a reinforcement learner leads to higher task success rate and reward in both simulations and against real users. We also present a fully neural end-to-end agent, trained entirely from user feedback, and discuss its application towards personalized dialogue agents. | Towards End-to-End Reinforcement Learning of Dialogue Agents for Information Access |
d18017180 | Recently, the development of neural machine translation (NMT) has significantly improved the translation quality of automatic machine translation. While most sentences are more accurate and fluent than translations by statistical machine translation (SMT)-based systems, in some cases, the NMT system produces translations that have a completely different meaning. This is especially the case when rare words occur.When using statistical machine translation, it has already been shown that significant gains can be achieved by simplifying the input in a preprocessing step. A commonly used example is the pre-reordering approach.In this work, we used phrase-based machine translation to pre-translate the input into the target language. Then a neural machine translation system generates the final hypothesis using the pre-translation. Thereby, we use either only the output of the phrase-based machine translation (PBMT) system or a combination of the PBMT output and the source sentence.We evaluate the technique on the English to German translation task. Using this approach we are able to outperform the PBMT system as well as the baseline neural MT system by up to 2 BLEU points. We analyzed the influence of the quality of the initial system on the final result. | Pre-Translation for Neural Machine Translation |
d258866084 | Building Natural Language Understanding (NLU) capabilities for Indic languages, which have a collective speaker base of more than one billion speakers is absolutely crucial. In this work, we aim to improve the NLU capabilities of Indic languages by making contributions along 3 important axes (i) monolingual corpora (ii) NLU testsets (iii) multilingual LLMs focusing on Indic languages. Specifically, we curate the largest monolingual corpora, IndicCorp, with 20.9B tokens covering 24 languages from 4 language families -a 2.3x increase over prior work, while supporting 12 additional languages. Next, we create a humansupervised benchmark, IndicXTREME, consisting of nine diverse NLU tasks covering 20 languages. Across languages and tasks, IndicX-TREME contains a total of 105 evaluation sets, of which 52 are new contributions to the literature. To the best of our knowledge, this is the first effort towards creating a standard benchmark for Indic languages that aims to test the multilingual zero-shot capabilities of pretrained language models. Finally, we train IndicBERT v2, a state-of-the-art model supporting all the languages. Averaged across languages and tasks, the model achieves an absolute improvement of 2 points over a strong baseline. The data and models are available at https:// github.com/AI4Bharat/IndicBERT. | Towards Leaving No Indic Language Behind: Building Monolingual Corpora, Benchmark and Models for Indic Languages |
d19935188 | Online topic modeling, i.e., topic modeling with stochastic variational inference, is a powerful and efficient technique for analyzing large datasets, and ADAGRAD is a widely-used technique for tuning learning rates during online gradient optimization. However, these two techniques do not work well together. We show that this is because ADAGRAD uses accumulation of previous gradients as the learning rates' denominators. For online topic modeling, the magnitude of gradients is very large. It causes learning rates to shrink very quickly, so the parameters cannot fully converge until the training ends. | Why ADAGRAD Fails for Online Topic Modeling |
d52939688 | In this paper, we propose a new rich resource enhanced AMR aligner which produces multiple alignments and a new transition system for AMR parsing along with its oracle parser. Our aligner is further tuned by our oracle parser via picking the alignment that leads to the highestscored achievable AMR graph. Experimental results show that our aligner outperforms the rule-based aligner in previous work by achieving higher alignment F1 score and consistently improving two open-sourced AMR parsers. Based on our aligner and transition system, we develop a transition-based AMR parser that parses a sentence into its AMR graph directly. An ensemble of our parsers with only words and POS tags as input leads to 68.4 Smatch F1 score, which outperforms the parser of Wang and Xue (2017). | An AMR Aligner Tuned by Transition-based Parser |
d1564849 | The centroid-based model for extractive document summarization is a simple and fast baseline that ranks sentences based on their similarity to a centroid vector. In this paper, we apply this ranking to possible summaries instead of sentences and use a simple greedy algorithm to find the best summary. Furthermore, we show possibilities to scale up to larger input document collections by selecting a small number of sentences from each document prior to constructing the summary. Experiments were done on the DUC2004 dataset for multi-document summarization. We observe a higher performance over the original model, on par with more complex state-of-the-art methods. | Revisiting the Centroid-based Method: A Strong Baseline for Multi-Document Summarization |
d4421747 | Semantic Textual Similarity (STS) measures the meaning similarity of sentences. Applications include machine translation (MT), summarization, generation, question answering (QA), short answer grading, semantic search, dialog and conversational systems. The STS shared task is a venue for assessing the current state-of-the-art. The 2017 task focuses on multilingual and cross-lingual pairs with one sub-track exploring MT quality estimation (MTQE) data. The task obtained strong participation from 31 teams, with 17 participating in all language tracks. We summarize performance and review a selection of well performing methods. Analysis highlights common errors, providing insight into the limitations of existing models. To support ongoing work on semantic representations, the STS Benchmark is introduced as a new shared training and evaluation set carefully selected from the corpus of English STS shared task data (2012)(2013)(2014)(2015)(2016)(2017). | SemEval-2017 Task 1: Semantic Textual Similarity Multilingual and Cross-lingual Focused Evaluation |
d51782450 | Edinburgh Research Explorer Unsupervised Source Hierarchies for Low-Resource Neural Machine Translation | |
d16261459 | Sentiments expressed in user-generated short text and sentences are nuanced by subtleties at lexical, syntactic, semantic and pragmatic levels. To address this, we propose to augment traditional features used for sentiment analysis and sarcasm detection, with cognitive features derived from the eye-movement patterns of readers. Statistical classification using our enhanced feature set improves the performance (F-score) of polarity detection by a maximum of 3.7% and 9.3% on two datasets, over the systems that use only traditional features. We perform feature significance analysis, and experiment on a held-out dataset, showing that cognitive features indeed empower sentiment analyzers to handle complex constructs. | Leveraging Cognitive Features for Sentiment Analysis |
d264038759 | La production d'annotations linguistiques ou gloses interlinéaires explicitant le sens ou la fonction de chaque unité repérée dans un enregistrement source (ou dans sa transcription) est une étape importante du processus de documentation des langues.Ces gloses exigent une très grande expertise de la langue documentée et un travail d'annotation fastidieux.Notre étude s'intéresse à l'automatisation partielle de ce processus.Il s'appuie sur la partition des gloses en deux types : les gloses grammaticales exprimant une fonction grammaticale et les gloses lexicales indiquant les unités de sens.Notre approche repose sur l'hypothèse d'un alignement entre les gloses lexicales et une traduction ainsi que l'utilisation de Lost, un modèle probabiliste de traduction automatique.Nos expériences sur une langue en cours de documentation, le tsez, montrent que cet apprentissage est effectif même avec un faible nombre de phrases de supervision. | Production automatique de gloses interlinéaires à travers un modèle probabiliste exploitant des alignements |
d264038774 | L'annotation manuelle de corpus est un processus coûteux et lent, notamment pour la tâche de reconnaissance d'entités nommées.L'apprentissage actif vise à rendre ce processus plus efficace, en sélectionnant les portions les plus pertinentes à annoter.Certaines stratégies visent à sélectionner les portions les plus représentatives du corpus, d'autres, les plus informatives au modèle de langage.Malgré un intérêt grandissant pour l'apprentissage actif, rares sont les études qui comparent ces différentes stratégies dans un contexte de reconnaissance d'entités nommées médicales.Nous proposons une comparaison de ces stratégies en fonction des performances de chacune sur 3 corpus de documents cliniques en langue française : MERLOT, QuaeroFrenchMed et E3C.Nous comparons les stratégies de sélection mais aussi les différentes façons de les évaluer.Enfin, nous identifions les stratégies qui semblent les plus efficaces et mesurons l'amélioration qu'elles présentent, à différentes phases de l'apprentissage. | Stratégies d'apprentissage actif pour la reconnaissance d'entités nommées en français |
d263850903 | The availability of language representations learned by large pretrained neural network models (such as BERT and ELECTRA) has led to improvements in many downstream Natural Language Processing tasks in recent years. Pretrained models usually differ in pretraining objectives, architectures, and datasets they are trained on which can affect downstream performance. In this contribution, we fine-tuned German BERT and German ELECTRA models to identify toxic (subtask 1), engaging (subtask 2), and fact-claiming comments (subtask 3) in Facebook data provided by the GermEval 2021 competition. We created ensembles of these models and investigated whether and how classification performance depends on the number of ensemble members and their composition. On out-of-sample data, our best ensemble achieved a macro-F1 score of 0.73 (for all subtasks), and F1 scores of 0.72, 0.70, and 0.76 for subtasks 1, 2, and 3, respectively. | FHAC at GermEval 2021: Identifying German toxic, engaging, and fact-claiming comments with ensemble learning |
d139787 | Existing knowledge-based question answering systems often rely on small annotated training data. While shallow methods like relation extraction are robust to data scarcity, they are less expressive than the deep meaning representation methods like semantic parsing, thereby failing at answering questions involving multiple constraints. Here we alleviate this problem by empowering a relation extraction method with additional evidence from Wikipedia. We first present a neural network based relation extractor to retrieve the candidate answers from Freebase, and then infer over Wikipedia to validate these answers. Experiments on the WebQuestions question answering dataset show that our method achieves an F 1 of 53.3%, a substantial improvement over the state-of-the-art. | Question Answering on Freebase via Relation Extraction and Textual Evidence |
d258987847 | Stance detection determines whether the author of a piece of text is in favor of, against, or neutral towards a specified target, and can be used to gain valuable insights into social media. The ubiquitous indirect referral of targets makes this task challenging, as it requires computational solutions to model semantic features and infer the corresponding implications from a literal statement. Moreover, the limited amount of available training data leads to subpar performance in out-of-domain and cross-target scenarios, as data-driven approaches are prone to rely on superficial and domain-specific features. In this work, we decompose the stance detection task from a linguistic perspective, and investigate key components and inference paths in this task. The stance triangle is a generic linguistic framework previously proposed to describe the fundamental ways people express their stance. We further expand it by characterizing the relationship between explicit and implicit objects. We then use the framework to extend one single training corpus with additional annotation. Experimental results show that strategically-enriched data can significantly improve the performance on out-of-domain and cross-target evaluation. | Guiding Computational Stance Detection with Expanded Stance Triangle Framework |
d30339746 | 1) LORIA -UMR 7503, 54506 Vandoeuvre-lès-Nancy Cedex (2) Université de Lorraine Slim.Ouni@loria.fr RESUME ____________________________________________________________________________________________________________ Dans cet article, nous présentons VisArtico, un logiciel de visualisation de données articulatoires obtenues par un articulographe, l'AG500. Ce logiciel permet de visualiser les positions des capteurs et de les animer simultanément avec l'acoustique : l'utilisateur a la possibilité de visualiser le contour de la langue et des lèvres. Il permet également de trouver le plan midsagittal du locuteur, et déduire la position du palais, si cette information est absente lors de l'acquisition. De plus, VisArtico offre la possibilité d'étiqueter phonétiquement les trajectoires. D'autres fonctionnalités sont également décrites. L'objectif est de fournir un outil efficace de visualisation de données articulatoires qui peut être utile à toute personne étudiant la production de la parole.ABSTRACT _________________________________________________________________________________________________________VisArtico : visualizing articulatory data acquired by an articulographIn this paper, we present VisArtico, visualization software of articulatory data acquired by an articulograph, AG500. The software allows displaying the positions of the sensors that are simultaneously played with the speech signal. It is possible to display the tongue contour and the lips contour. The software helps to find the midsagittal plane of the speaker and find the palate contour. In addition, VisArtico allows labeling phonetically the articulatory data. Our main goal is to provide an efficient tool to visualize articulatory data for researchers working in speech production field. MOTS-CLES : Articulographe, visualisation, production de la parole, conduit vocal, EMA. | VisArtico : visualiser les données articulatoires obtenues par un articulographe |
d257687773 | In Task Oriented Dialogue (TOD) system, detecting and inducing new intents are two main challenges to apply the system in the real world. In this paper, we suggest the semantic multiview model to resolve these two challenges: (1) SBERT for General Embedding (GE), (2) Multi Domain Batch (MDB) for dialogue domain knowledge, and (3) Proxy Gradient Transfer (PGT) for cluster-specialized semantic. MDB feeds diverse dialogue datasets to the model at once to tackle the multi-domain problem by learning the multiple domain knowledge. We introduce a novel method PGT, which employs the Siamese network to fine-tune the model with a clustering method directly. Our model can learn how to cluster dialogue utterances by using PGT. Experimental results demonstrate that our multi-view model with MDB and PGT significantly improves the Open Intent Induction performance compared to baseline systems. | Multi-View Zero-Shot Open Intent Induction from Dialogues: Multi Domain Batch and Proxy Gradient Transfer |
d102353364 | Various NLP problems -such as the prediction of sentence similarity, entailment, and discourse relations -are all instances of the same general task: the modeling of semantic relations between a pair of textual elements. A popular model for such problems is to embed sentences into fixed size vectors, and use composition functions (e.g. concatenation or sum) of those vectors as features for the prediction. At the same time, composition of embeddings has been a main focus within the field of Statistical Relational Learning (SRL) whose goal is to predict relations between entities (typically from knowledge base triples). In this article, we show that previous work on relation prediction between texts implicitly uses compositions from baseline SRL models. We show that such compositions are not expressive enough for several tasks (e.g. natural language inference). We build on recent SRL models to address textual relational problems, showing that they are more expressive, and can alleviate issues from simpler compositions. The resulting models significantly improve the state of the art in both transferable sentence representation learning and relation prediction. | Composition of Sentence Embeddings: Lessons from Statistical Relational Learning |
d258378210 | Language label tokens are often used in multilingual neural language modeling and sequence-to-sequence learning to enhance the performance of such models. An additional product of the technique is that the models learn representations of the language tokens, which in turn reflect the relationships between the languages. In this paper, we study the learned representations of dialects produced by neural dialect-to-standard normalization models. We use two large datasets of typologically different languages, namely Finnish and Norwegian, and evaluate the learned representations against traditional dialect divisions of both languages. We find that the inferred dialect embeddings correlate well with the traditional dialects. The methodology could be further used in noisier settings to find new insights into language variation. | Dialect Representation Learning with Neural Dialect-to-Standard Normalization |
d202758970 | This paper presents a syntactic treebank for spoken Naija, an English pidgincreole, which is rapidly spreading across Nigeria. The syntactic annotation is developed in the Surface-Syntactic Universal Dependency annotation scheme (SUD) (Gerdes et al., 2018) and automatically converted into UD. We present the workflow of the treebank development for this under-resourced language. A crucial step in the syntactic analysis of a spoken language consists in manually adding a markup onto the transcription, indicating the segmentation into major syntactic units and their internal structure. We show that this so-called "macrosyntactic" markup improves parsing results. We also study some iconic syntactic phenomena that clearly distinguish Naija from English. 1 ISLE Meta Data Initiative (IMDI) is a metadata standard to describe multi-media and multi-modal language resources. 2 https://tla.mpi.nl/tools/tla-tools/arbil/ | A Surface-Syntactic UD Treebank for Naija |
d239998725 | Word Sense Disambiguation (WSD) aims to automatically identify the exact meaning of one word according to its context. Existing supervised models struggle to make correct predictions on rare word senses due to limited training data and can only select the best definition sentence from one predefined word sense inventory (e.g., WordNet). To address the data sparsity problem and generalize the model to be independent of one predefined inventory, we propose a gloss alignment algorithm that can align definition sentences (glosses) with the same meaning from different sense inventories to collect rich lexical knowledge. We then train a model to identify semantic equivalence between a target word in context and one of its glosses using these aligned inventories, which exhibits strong transfer capability to many WSD tasks 1 . Experiments on benchmark datasets show that the proposed method improves predictions on both frequent and rare word senses, outperforming prior work by 1.2% on the All-Words WSD Task and 4.3% on the Low-Shot WSD Task. Evaluation on WiC Task also indicates that our method can better capture word meanings in context. | Connect-the-Dots: Bridging Semantics between Words and Definitions via Aligning Word Sense Inventories |
d52957274 | Language Models (LMs) are important components in several Natural Language Processing systems. Recurrent Neural Network LMs composed of LSTM units, especially those augmented with an external memory, have achieved state-of-the-art results. However, these models still struggle to process long sequences which are more likely to contain longdistance dependencies because of information fading and a bias towards more recent information. In this paper we demonstrate an effective mechanism for retrieving information in a memory augmented LSTM LM based on attending to information in memory in proportion to the number of timesteps the LSTM gating mechanism persisted the information. | Persistence pays off: Paying Attention to What the LSTM Gating Mechanism Persists |
d259376809 | This paper describes the system we used to participate in the shared task(Kiesel et al., 2023), as well as additional experiments beyond the scope of the shared task, but using its data. Our primary goal is to compare the effectiveness of transformers model compared to low-resource dictionaries. Secondly, we compare the difference in performance of a learned dictionary and of a dictionary designed by experts in the field of values. Our findings surprisingly show that transformers perform on par with a dictionary containing less than 1k words, when evaluated with 19 fine-grained categories, and only outperform a dictionary-based approach in a coarse setting with 10 categories. Interestingly, the expert dictionary has a precision on par with the learned one, while its recall is clearly lower, potentially an indication of overfitting of topics to values in the shared task's dataset. Our findings should be of interest to both the NLP and Value scientific communities on the use of automated approaches for value classification. | TeamEC at SemEval-2023 Task 4: Transformers VS. Low-Resource Dictionaries, Expert Dictionary VS. Learned Dictionary |
d259376838 | Our team silp_nlp participated in SemEval2023 Task 2: MultiCoNER II. Our work made systems for 11 mono-lingual tracks. For leveraging the advantage of all track knowledge we chose transformer-based pretrained models, which have strong cross-lingual transferability. Hence our model trained in two stages, the first stage for multi-lingual learning from all tracks and the second for fine-tuning individual tracks. Our work highlights that the knowledge of all tracks can be transferred to an individual track if the baseline language model has crosslingual features. Our system positioned itself in the top 10 for 4 tracks by scoring 0.7432 macro F1 score for the Hindi track ( 7th rank ) and 0.7322 macro F1 score for the Bangla track ( 9th rank ). | silp_nlp at SemEval-2023 Task 2: Cross-lingual Knowledge Transfer for Mono-lingual Learning |
d253628204 | Image captioning is a prominent Artificial Intelligence (AI) research area that deals with visual recognition and a linguistic description of the image. It is an interdisciplinary field concerning how computers can see and understand digital images & videos, and describe them in a language known to humans. Constructing a meaningful sentence needs both structural and semantic information of the language. This paper highlights the contribution of image caption generation for the Assamese language. The unavailability of an image caption generation system for the Assamese language is an open problem for AI-NLP researchers, and it's just an early stage of the research. To achieve our defined objective, we have used the encoder-decoder framework, which combines the Convolutional Neural Networks and the Recurrent Neural Networks. The experiment has been tested on Flickr30k and Coco Captions dataset, which have been originally present in the English language. We have translated these datasets into Assamese language using the state-of-the-art Machine Translation (MT) system for our designed work. | Image Caption Generation for Low-Resource Assamese Language |
d10479248 | We introduce an LSTM-based method for dynamically integrating several wordprediction experts to obtain a conditional language model which can be good simultaneously at several subtasks. We illustrate this general approach with an application to dialogue where we integrate a neural chat model, good at conversational aspects, with a neural question-answering model, good at retrieving precise information from a knowledge-base, and show how the integration combines the strengths of the independent components. We hope that this focused contribution will attract attention on the benefits of using such mixtures of experts in NLP. 1 | LSTM-based Mixture-of-Experts for Knowledge-Aware Dialogues |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.