Spain

&quot; Pre-training large-scale language models (LMs) requires huge amounts of text corpora. 
 LMs for English enjoy the ever-growing corpora of diverse language resources.
 However, less-resourced languages and their mono- and multilingual LMs often struggle to obtain bigger datasets.
  A typical approach in this case implies using machine translation of English corpora to a target language.
  In this work, we study the caveats of applying directly translated corpora for fine-tuning LMs for downstream natural language processing tasks and demonstrate that careful curation along with post-processing leads to improved performance and overall LMs robustness.&quot;

ECAI 2020

How Machine Translated Corpora Affects Language Models in MRC Tasks

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-23 is the Thirty-Seventh AAAI Conference on Artificial Intelligence. The theme of this conference is to create collaborative bridges within and beyond AI. Like previous AAAI conferences, AAAI-23 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and two new activities: a Bridge Program and a Lab Program. Many of these activities are tailored to the theme of bridges and all are selected according to the highest standards, with additional programs for students and young researchers. 
AAAI is providing you with a conference planner, which you can use to help organize your itinerary of activities. This includes talks to attend in person, talks to attend remotely, breaks with colleagues and your site seeing activities. To access this conference planner, please go to [https://aaai-2023.takemobi.io](https://aaai-2023.takemobi.io).

In order to access this site, you need to register. If you haven't already, please register [here](https://aaai.org/Conferences/AAAI-23/registration/).


AAAI 2023

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines.

" Pre-training large-scale language models (LMs) requires huge amounts of text corpora. 
 LMs for English enjoy the ever-growing corpora of diverse language resources.
 However, less-resourced languages and their mono- and multilingual LMs often struggle to obtain bigger datasets.
  A typical approach in this case implies using machine translation of English corpora to a target language.
  In this work, we study the caveats of applying directly translated corpora for fine-tuning LMs for downstream natural language processing tasks and demonstrate that careful curation along with post-processing leads to improved performance and overall LMs robustness."

technical paper

ECAI2020 featured a full programme of technical papers, as well as many AI-related events of interest for researchers, students and all attendants who are interested on contemporary AI, including:




**An extensive programme of workshops, tutorials, and invited speakers**

**The Prestigious Applications of Intelligent Systems conference (PAIS)**

**The Frontiers in AI track sessions**

**The Lab To Market event**

**The Women in AI time**

**The EU Challenges forum**

… and many more!

ECAI 2020 featured a full programme of technical papers, as well as many AI-related events of interest for researchers, students and all attendants who are interested on contemporary AI.

Structural Onomatology for Username Generation: A Partial Account

Applying LSTM Networks to Predict Multi-drug Resistance Using Binary Multivariate Clinical Sequences

Modelling an Intelligent Tutoring System based on Supervised Behaviour Biometrics

On Multi-Agent Coordination of Agri-Robot Fleets

I will briefly survey recent and expected developments in AI and their implications. Some are enormously positive, while others, such as the development of autonomous weapons and the replacement of humans in economic roles, may be negative. Beyond these, one must expect that AI capabilities will eventually exceed those of humans across a range of real-world-decision making scenarios. Should this be a cause for concern, as Elon Musk, Stephen Hawking, and others have suggested? And, if so, what can we do about it? While some in the mainstream AI community dismiss the issue, I will argue that the problem is real and that the technical aspects of it are solvable if we replace current definitions of AI with a version based on provable benefit to humans.

How not to destroy the world with AI

Earth observation sensors deliver ever-expanding collections of geospatial data at multiple resolutions (spatial, temporal and thematic or spectral). Efficient tools to extract knowledge from these collections are currently missing. Here we present the first release of Clustering geo-Data Cubes (CDC), a Python package to cluster geospatial data cubes by explicitly considering their dimensionality. CDC has three main hallmarks: 1/ it is based on innovative co- and tri-clustering methods that identify groups of pixels with similar spatio-temporal and/or thematic information by simultaneously considering all the dimensions of the data. This overcomes a major limitations of traditional clustering approaches, which analyze each dimension separately; 2/ it provides refined clusters by re-grouping the results obtained from co- and/or tri-clustering. These refined clusters better capture the patterns present in the data and represent a more automatic approach to analyze geospatial data cubes because the number of clusters is automatically chosen via an optimization procedure; and 3/ it allows users to run tasks efficiently by either using NumPy’s threading capabilities or Dask’s parallel computing power. Hence, CDC is a scalable package that can analyze both small and big geospatial data cubes. These hallmarks are showcased through several case studies

Clustering geo-data cubes

The prosumers flexibility procurement has increased due to the current penetration of distributed and variable renewable energy sources. The prosumers flexibility is often able to quickly adjust the power consumption, making it well suited as a primary and secondary reserve for ancillary services. In the era of smart grids, the role of the aggregator has been increasingly exploited and considered as a player that can facilitate small prosumers’ participation in electricity markets. This paper proposes an approach based on the use of prosumers flexibility by an aggregator to support ancillary services at a low voltage level. An asymmetric pool market approach is considered for flexibility negotiation between prosumers and the local marker operator (aggregator). From the achieved results it is possible to conclude that the use of flexibility can bring technical and economic benefits for network operators.

Prosumers Flexibility as Support for Ancillary Services in Low Voltage Level

"A constant stream of plasma is continuously flowing from the Sun. This solar wind reaches the Earth and can have serious effects on our technological infrastructure and health. Multiple spacecraft to measure this plasma and its properties. Understanding where in the Sun such plasma originates help us to understand the connection between our planet and our star. We use modern machine learning techniques that extract automatically data from millions of hours of spacecraft measurements. These techniques do not require the intervention of human experts. Our goal is to allow the computer to discover the properties of the solar plasma without the bias of the scientist. We show in our work how this can be achieved and how it helps the scientists to make new discoveries.

We will present one example of such techniques: unsupervised clustering can be used to segregate different types of solar wind. We propose the use of advanced data reduction methods to process the data, and we introduce the use of multiple classification methods to visualize and interpret 14 years of spacecraft data. Finally, we show how these techniques can potentially be used to uncover hidden information, and how they compare with previous empirical methods."

Machine Learning in Heliophysical Applications: the Example of Unsupervised Solar Wind Classification

This talk focuses on the concept of autonomy and describes a research path followed over the years  on the design and implementation of  software architectures for controlling complex subsystems operating in space, for example an exploration rover.  The leading thread of the talk is the presentation of innovative ideas to increase the level of  autonomy for an intelligent controller. Starting from the description of a mixed-initiative system developed within the ESA Mars-Express mission in operation at ESA-ESOC since 2005, the talk moves on to presenting a new interpretation of the known Sense-Plan-Act (SPA) autonomous and deliberation paradigm and some correlated software architectures. SPA is one of the core ideas to realize autonomous systems, and AI Planning & Scheduling, as well as their integration, are among the basic technologies it is based upon.  Then, an extended paradigm, named Discover-Plan-Act is described, where state-of-the-art AI planning systems are integrated with Reinforcement Learning (RL) algorithms guided by intrinsic motivations (curiosity, exploration, novelty, surprise). A discussion about the future work and the open research challenges closes the talk.

Extending the autonomy envelope of space applications: a research path

In this presentation I give an overview of the progress in H2020 project ExtremeEarth (http://earthanalytics.eu/) which investigates scalable big data and deep learning techniques for Copernicus data. The project is driven by two use cases in the areas of Food Security and Polar Regions.The former use case aims to develop irrigation recommendations from farmers while the latter will produce ice maps that can aid maritime users in the Arctic. The contributions of the project in the area of deep learning are twofold: (i) the production of two very large training datasets (millions of samples) for crop type mapping and ice classification, and (ii) the study of LSTM and CNN architectures for the same problems. In the area of big data, the project is developing four big linked geospatial data systems and a benchmarking environment for evaluating their performance and scalability. The development of all algorithms, systems and use cases takes place in the big data and AI platform Hopsworks (https://www.logicalclocks.com/) and run on CREODIAS.

How Machine Translated Corpora Affects Language Models in MRC Tasks

Downloads

Next from ECAI 2020

Structural Onomatology for Username Generation: A Partial Account

Similar lecture

The Effect of Modeling Human Rationality Level on Learning Rewards from Multiple Feedback Types

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

How Machine Translated Corpora Affects Language Models in MRC Tasks

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from ECAI 2020

Structural Onomatology for Username Generation: A Partial Account

Similar lecture

The Effect of Modeling Human Rationality Level on Learning Rewards from Multiple Feedback Types

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads