Events

This workshop introduces the basic concepts of Deep Learning — the training and performance evaluation of large neural networks, especially for image classification, natural language processing, and time-series data. Like many other machine learning algorithms, we will use deep learning algorithms to map input data to their appropriately classified outcome labels. You will use the R interface to Keras to become familiar with basic concepts like input and output layers, batch sizes and output dimensions, dropout rates, weight parametrization and bias, backpropagation, and loss, activation, and optimization functions. You will also gain confidence exploring more complex approaches that utilize pretrained and fine-tuned models. Register: https://dlab.berkeley.edu/events/r-deep-learning-parts-1-2/2023-04-19 Prerequisites: D-Lab's Intro to Machine Learning in R workshop series or equivalent introductory machine learning knowledge. Workshop Materials: https://github.com/dlab-berkeley/R-Deep-Learning(link is external) Software Requirements:Installation Instructions(link is external) for R and RStudio Is RStudio Not working on your laptop?Attend the workshop anyway, we can provide you with a cloud-based solution(link is external) until you figure out the problems with your local installation.

Join us for the Map Mashup event kickoff and OpenGeo Lightning Talks, including choosing a map, cartography, georeferencing, open source tools, open data and more! Beverages & snacks will be provided. REGISTER in advance to receive a Zoom link for this hybrid event. Location: Earth Sciences & Map Library Seminar Room (55A McCone Hall) & Zoom This event is part of Map Mashup, a two-day map-making extravaganza. Anyone is welcome to participate in this event, even if you do not plan on submitting a Map Mashup creation. The UC Berkeley Earth Sciences & Map Library invites students and others to use maps from our collection as the basis for a new map that they create. Over the course of two days we're hosting a series of map-related workshops and open work hours for people to work on their maps, either individually or in groups. The map mashups can be created on a computer or using physical materials. We'll publish digital versions of the final map mashups online!