informatik-vortraege August 2021

informatik-vortraege@lists.rwth-aachen.de

3 participants
4 discussions

Einladung Informatik-Oberseminar Malte Nuhn

by Sekretariat I6

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Freitag, 12. Juli 2019, 10.00 Uhr Ort: Informatikzentrum, E3, Raum 9222 Referent: Dipl.-Inform. Malte Nuhn Thema: Unsupervised Training with Applications in Natural Language Processing// Abstract: The state-of-the-art algorithms for various natural language processing tasks require large amounts of labeled training data. At the same time, obtaining labeled data of high quality is often the most costly step in setting up natural language processing systems.Opposed to this, unlabeled data is much cheaper to obtain and available in larger amounts.Currently, only few training algorithms make use of unlabeled data. In practice, training with only unlabeled data is not performed at all. In this thesis, we study how unlabeled data can be used to train a variety of models used in natural language processing. In particular, we study models applicable to solving substitution ciphers, spelling correction, and machine translation. This thesis lays the groundwork for unsupervised training by presenting and analyzing the corresponding models and unsupervised training problems in a consistent manner.We show that the unsupervised training problem that occurs when breaking one-to-one substitution ciphers is equivalent to the quadratic assignment problem (QAP) if a bigram language model is incorporated and therefore NP-hard. Based on this analysis, we present an effective algorithm for unsupervised training for deterministic substitutions. In the case of English one-to-one substitution ciphers, we show that our novel algorithm achieves results close to human performance, as presented in [Shannon 49]. Also, with this algorithm, we present, to the best of our knowledge, the first automatic decipherment of the second part of the Beale ciphers.Further, for the task of spelling correction, we work out the details of the EM algorithm [Dempster & Laird + 77] and experimentally show that the error rates achieved using purely unsupervised training reach those of supervised training.For handling large vocabularies, we introduce a novel model initialization as well as multiple training procedures that significantly speed up training without hurting the performance of the resulting models significantly.By incorporating an alignment model, we further extend this model such that it can be applied to the task of machine translation. We show that the true lexical and alignment model parameters can be learned without any labeled data: We experimentally show that the corresponding likelihood function attains its maximum for the true model parameters if a sufficient amount of unlabeled data is available. Further, for the problem of spelling correction with symbol substitutions and local swaps, we also show experimentally that the performance achieved with purely unsupervised EM training reaches that of supervised training. Finally, using the methods developed in this thesis, we present results on an unsupervised training task for machine translation with a ten times larger vocabulary than that of tasks investigated in previous work. Es laden ein: die Dozentinnen und Dozenten der Informatik _______________________________________________ -- -- Stephanie Jansen Faculty of Mathematics, Computer Science and Natural Sciences HLTPR - Human Language Technology and Pattern Recognition RWTH Aachen University Ahornstraße 55 D-52074 Aachen Tel. Frau Jansen: +49 241 80-216 06 Tel. Frau Andersen: +49 241 80-216 01 Fax: +49 241 80-22219 sek(a)i6.informatik.rwth-aachen.de www.hltpr.rwth-aachen.de Tel: +49 241 80-216 01/06 Fax: +49 241 80-22219 sek(a)i6.informatik.rwth-aachen.de www.hltpr.rwth-aachen.de

2 years, 3 months

Einladung: Informatik-Oberseminar Vipin Ravindran Vijayalakshmi

by Bolke-Hermanns, Helene

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Montag, 6. September 2021, 10:00 Uhr Zoom: https://rwth.zoom.us/j/91549249969?pwd=K1VGbDhkaUlpUDdhSFR3dXh1aWFzQT09 Meeting ID: 915 4924 9969 Passcode: 391915 Referent: Vipin Ravindran Vijayalakshmi, M. Sc. Lehrstuhl für Management Science Thema: Selfishness in Strategic Resource Allocation Problems Abstract: Non-cooperative game theory has emerged as an essential tool in analyzing and predicting the outcome of decentralized systems. Various algorithmic aspects arising due to strategic behavior among multiple non-cooperative users in several classes of resource allocation problems can be studied using a game theoretic approach. In this talk, we consider congestion games which are often used to model various scenarios of resource allocation by non-cooperative users. Congestion games constitute of a class of games in which a pure Nash equilibrium always exists. The hardness of computing a pure Nash equilibrium in congestion games has been of significant interest in the scientific community over the past two decades. As computing an exact pure Nash equilibrium is known to be hard, we study a weaker notion of pure Nash equilibrium called an approximate pure Nash equilibrium and to that extend, analyze an efficient algorithm that computes approximate pure Nash equilibria in congestion games with an improved approximation guarantee. In spite of being the predominant class of games to model resource allocation problems involving different users, congestion games lack an element of time dependence, especially in certain application scenarios such as the road transportation network. It is quite unnatural to assume that all users of a road section experience the same congestion. We consider a scheduling game on parallel machines, in which jobs try to minimize their completion time by choosing a machine to be processed on. Each machine uses an individual priority list to decide on the order according to which the jobs on the machine are processed. Here, we study the existence of a pure Nash equilibrium and characterize classes of instances in which a pure Nash equilibrium is guaranteed to exist. Es laden ein: die Dozentinnen und Dozenten der Informatik

3 years

Einladung: Informatik-Oberseminar Peter Lindner

by Peter Lindner

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Dienstag, 24. August 2021, 10:00 Uhr Zoom: https://rwth.zoom.us/j/97621241861?pwd=TURveUs0YVRWck5XRWprNEJCdFJmZz09 Meeting ID: 976 2124 1861 Passcode: 592961 Referent: Peter Lindner, M. Sc. Lehrstuhl für Informatik 7 (Logik und Theorie diskreter Systeme) Thema: The Theory of Infinite Probabilistic Databases Abstract: Probabilistic (relational) databases extend the conventional relational database model by probability distributions over database instances. Such models are desirable for, and applicable to, a wide range of scenarios where data is subject to uncertainty. For a long time, the theoretical state of the art was to view a probabilistic database as a discrete probability space with only finitely many possible outcome instances. In this talk, we give an overview over our contributions regarding the extension of this model, along with various key concepts, to infinite probability spaces. This is required to support basic probabilistic methods, like continuous noise models and tackles default shortcomings of the restriction to discrete spaces. To this end, we present and explore a suitable unifying framework for infinite probabilistic databases that is based on point processes. This covers all applications of interest, and complies with the desired requirements for such a framework. It's inherent abstractness allows us to discuss central notions from the theory of probabilistic databases like query semantics and independence assumptions from a plain probability theoretic point of view. Es laden ein: die Dozentinnen und Dozenten der Informatik

3 years

Einladung: Informatik-Oberserminar Francis Engelmann

by Francis Engelmann

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Montag, 16. August 2021, 14.00 Uhr Ort: Zoom Videokonferenz https://rwth.zoom.us/j/97522924662?pwd=NWN1Z2lkWGJOTkt0N2QyTHBRdSs1Zz09 <https://www.google.com/url?q=https://rwth.zoom.us/j/97522924662?pwd%3DNWN1Z…> Meeting ID: 975 2292 4662 Passcode: 451032 Referent: Francis Engelmann, M.Sc. Lehrstuhl Informatik 13 Thema: 3D Scene Understanding on Point Clouds Abstract: In this talk I present my thesis contributions to the emerging field of 3D scene understanding. That is, given a 3D scene representation as input, we address tasks such as 3D object detection, shape reconstruction and pose estimation, as well as 3D semantic- and instance-segmentation. The recent availability of inexpensive depth sensors has made 3D data widely accessible. At the same time, current aspirations in the field of robotics, augmented reality and self-driving cars require efficient and reliable algorithms for understanding different 3D scene representations, such as polygon meshes, point clouds or volumetric structures. While 3D data overcomes inherent limitations of projected 2D views, such as occlusions, scale-ambiguity and lack of geometry, it also introduces new challenges including sparsity and non-uniform sampling. Therefore, existing methods for 2D image processing do not generalize well to 3D data structures. In this talk, we present novel approaches specific to 3D scene understanding. The main contributions are organized into three parts: The core contribution of the first part is a probabilistic formulation which integrates 3D shape and motion priors as well as stereo depth measurements into a global optimization problem. The resulting approach can jointly estimate the 3D shape, pose and motion of multiple vehicles in urban street scenes. The second part deals with new deep learning models for processing 3D point clouds. In particular, we propose sequential and recurrent consolidation units for increasing the spatial context of point networks, and a simple yet efficient dilation mechanism for increasing the receptive field size of deep point convolutional networks. Finally, in the third part, we introduce advanced deep learning models. For semantic segmentation, we present the combination of two types of convolutions operating jointly on point clouds and mesh surfaces. In instance segmentation, we propose a new paradigm combining bottom-up and top-down approaches as introduced in previous works. The talk concludes with a discussion and directions for future research.

3 years, 1 month

2024

2023

2022

2021

2020

2019

2018

informatik-vortraege August 2021