004 Datenverarbeitung; Informatik
Refine
Year of publication
Document Type
- Conference Proceeding (53) (remove)
Has Fulltext
- yes (53)
Is part of the Bibliography
- no (53)
Keywords
Institute
- Informatik (33)
- Wirtschaftswissenschaften (7)
- Informatik und Mathematik (4)
- Universitätsbibliothek (4)
- Physik (3)
- Center for Scientific Computing (CSC) (1)
- Geschichtswissenschaften (1)
- Präsidium (1)
The recognition of pharmacological substances, compounds and proteins is an essential preliminary work for the recognition of relations between chemicals and other biomedically relevant units. In this paper, we describe an approach to Task 1 of the PharmaCoNER Challenge, which involves the recognition of mentions of chemicals and drugs in Spanish medical texts. We train a state-of-the-art BiLSTM-CRF sequence tagger with stacked Pooled Contextualized Embeddings, word and sub-word embeddings using the open-source framework FLAIR. We present a new corpus composed of articles and papers from Spanish health science journals, termed the Spanish Health Corpus, and use it to train domain-specific embeddings which we incorporate in our model training. We achieve a result of 89.76% F1-score using pre-trained embeddings and are able to improve these results to 90.52% F1-score using specialized embeddings.
Despite the great importance of the Latin language in the past, there are relatively few resources available today to develop modern NLP tools for this language. Therefore, the EvaLatin Shared Task for Lemmatization and Part-of-Speech (POS) tagging was published in the LT4HALA workshop. In our work, we dealt with the second EvaLatin task, that is, POS tagging. Since most of the available Latin word embeddings were trained on either few or inaccurate data, we trained several embeddings on better data in the first step. Based on these embeddings, we trained several state-of-the-art taggers and used them as input for an ensemble classifier called LSTMVoter. We were able to achieve the best results for both the cross-genre and the cross-time task (90.64% and 87.00%) without using additional annotated data (closed modality). In the meantime, we further improved the system and achieved even better results (96.91% on classical, 90.87% on cross-genre and 87.35% on cross-time).
Retiming is a widely investigated technique for performance optimization. In general, it performs extensive modifications on a circuit netlist, leaving it unclear, whether the achieved performance improvement will still be valid after placement has been performed. This paper presents an approach for integrating retiming into a timing-driven placement environment. The experimental results show the benefit of the proposed approach on circuit performance in comparison with design flows using retiming only as a pre- or postplacement optimization method.
We study threshold testing, an elementary probing model with the goal to choose a large value out of n i.i.d. random variables. An algorithm can test each variable X_i once for some threshold t_i, and the test returns binary feedback whether X_i ≥ t_i or not. Thresholds can be chosen adaptively or non-adaptively by the algorithm. Given the results for the tests of each variable, we then select the variable with highest conditional expectation. We compare the expected value obtained by the testing algorithm with expected maximum of the variables. Threshold testing is a semi-online variant of the gambler’s problem and prophet inequalities. Indeed, the optimal performance of non-adaptive algorithms for threshold testing is governed by the standard i.i.d. prophet inequality of approximately 0.745 + o(1) as n → ∞. We show how adaptive algorithms can significantly improve upon this ratio. Our adaptive testing strategy guarantees a competitive ratio of at least 0.869 - o(1). Moreover, we show that there are distributions that admit only a constant ratio c < 1, even when n → ∞. Finally, when each box can be tested multiple times (with n tests in total), we design an algorithm that achieves a ratio of 1 - o(1).
In the last decade, much effort went into the design of robust third-person pronominal anaphor resolution algorithms. Typical approaches are reported to achieve an accuracy of 60-85%. Recent research addresses the question of how to deal with the remaining difficult-toresolve anaphors. Lappin (2004) proposes a sequenced model of anaphor resolution according to which a cascade of processing modules employing knowledge and inferencing techniques of increasing complexity should be applied. The individual modules should only deal with and, hence, recognize the subset of anaphors for which they are competent. It will be shown that the problem of focusing on the competence cases is equivalent to the problem of giving precision precedence over recall. Three systems for high precision robust knowledge-poor anaphor resolution will be designed and compared: a ruleset-based approach, a salience threshold approach, and a machine-learning-based approach. According to corpus-based evaluation, there is no unique best approach. Which approach scores highest depends upon type of pronominal anaphor as well as upon text genre.
This paper proposes a new approach for the encoding of images by only a few important components. Classically, this is done by the Principal Component Analysis (PCA). Recently, the Independent Component Analysis (ICA) has found strong interest in the neural network community. Applied to images, we aim for the most important source patterns with the highest occurrence probability or highest information called principal independent components (PIC). For the example of a synthetic image composed by characters this idea selects the salient ones. For natural images it does not lead to an acceptable reproduction error since no a-priori probabilities can be computed. Combining the traditional principal component criteria of PCA with the independence property of ICA we obtain a better encoding. It turns out that this definition of PIC implements the classical demand of Shannon’s rate distortion theory.
The annotation of texts and other material in the field of digital humanities and Natural Language Processing (NLP) is a common task of research projects. At the same time, the annotation of corpora is certainly the most time- and cost-intensive component in research projects and often requires a high level of expertise according to the research interest. However, for the annotation of texts, a wide range of tools is available, both for automatic and manual annotation. Since the automatic pre-processing methods are not error-free and there is an increasing demand for the generation of training data, also with regard to machine learning, suitable annotation tools are required. This paper defines criteria of flexibility and efficiency of complex annotations for the assessment of existing annotation tools. To extend this list of tools, the paper describes TextAnnotator, a browser-based, multi-annotation system, which has been developed to perform platform-independent multimodal annotations and annotate complex textual structures. The paper illustrates the current state of development of TextAnnotator and demonstrates its ability to evaluate annotation quality (inter-annotator agreement) at runtime. In addition, it will be shown how annotations of different users can be performed simultaneously and collaboratively on the same document from different platforms using UIMA as the basis for annotation.
The correspondence between the terminology used for querying and the one used in content objects to be retrieved, is a crucial prerequisite for effective retrieval technology. However, as terminology is evolving over time, a growing gap opens up between older documents in (long-term) archives and the active language used for querying such archives. Thus, technologies for detecting and systematically handling terminology evolution are required to ensure "semantic" accessibility of (Web) archive content on the long run. As a starting point for dealing with terminology evolution this paper formalizes the problem and discusses issues, first ideas and relevant technologies.
Interest to become a data scientist or related professions in data science domain is rapidly growing. To meet such a demand, we propose a novel educational service that aims to provide tailored learning paths for data science. Our target user is one who aims to be an expert in data science. Our approach is to analyze the background of the practitioner and match the learning units. A critical feature is that we use gamification to reinforce the practitioner engagement. We believe that our work provides a practical guideline for those who want to learn data science.
This paper shows the equivalence of applicative similarity and contextual approximation, and hence also of bisimilarity and contextual equivalence, in the deterministic call-by-need lambda calculus with letrec. Bisimilarity simplifies equivalence proofs in the calculus and opens a way for more convenient correctness proofs for program transformations. Although this property may be a natural one to expect, to the best of our knowledge, this paper is the first one providing a proof. The proof technique is to transfer the contextual approximation into Abramsky’s lazy lambda calculus by a fully abstract and surjective translation. This also shows that the natural embedding of Abramsky’s lazy lambda calculus into the call-by-need lambda calculus with letrec is an isomorphism between the respective term-models. We show that the equivalence property proven in this paper transfers to a call-by-need letrec calculus developed by Ariola and Felleisen. 1998 ACM Subject Classification: F.4.2, F.3.2, F.3.3, F.4.1. Key words and phrases: semantics, contextual equivalence, bisimulation, lambda calculus, call-by-need, letrec.