While deep ML models have been designed and operated mostly in the cloud, where there is enough computation power for their needs, we are starting to see a proliferation of ML models at the edge, on lower power devices. Increasing privacy awareness and latency are driving this trend of consuming data closer to its source.
In this talk, Valentin will focus on hybrid inference through splitting the model computations between the edge and the cloud. We ensure user data privacy by running the early computations of the model with raw data only on the edge. We will explore solutions for model inference but also training between the edge and cloud.
On March 15th, we're gathering for an introductory discussion to automatic differentiation where Bogdan will be talking about what automatic differentiation is, provide some examples and implememntation details.
Machine learning has become an indispensable part of modern society and continues to have a significant societal, economic and scientific impact.
On June 24th, we'll be hosting the third edition of First steps in AI, an event for all AI passionates that would like to study or continue studying artificial intelligence in one of the available academic programs from Iași.
For this event, our guests from Faculty of Computer Science, Mathematics, Automatic Control and Computer Engineering and Faculty of Electronics, Telecommunications and Information Technology will present the academic programs and courses that cover specific #AI topics in the upcoming academic year.
On June 14th we'll celebrate 5 years since our first meetup, back in 2017.
Join us at Acaju Cafe between 18:00 - 20:00 for a drink and some great conversations, along with community members and friends.
John is an international technology executive with over 35 years of experience in data and advanced analytics fields and has published two best-selling books on analytics and analytics teams, and has a 3rd book on the way.
John was an Executive Partner at Gartner, where he was management consultant to market leading companies in the areas of digital transformation, data monetization and advanced analytics. Before Gartner, he was responsible for the advanced analytics business unit of the Dell Software Group.
Scientific Machine Learning (SciML) represents a blend of scientific computing and machine learning methodologies.
Traditionally, scientific computing relies on mechanistic modeling (ie., differential equations), with a focus on domain models based on physical laws and scientific knowledge. However, while mechanistic models of lots of scientific phenomena are routinely available, our computational capabilities are unable to keep up with the increasingly demanding requirements of realistic simulation.
In December, we continue AI Global Festival series with our second meetup on predicting building temperature using artificial intelligence.
Alexandru Lungu, Senior Software Engineer at E.ON Software Development will share a practical use-case on applying AI for predicting building inside temperature, optimize power comsumption and make life more sustainable.
AI is not just a technology but an enabler for business value creation. Hence, AI is increasingly important as a topic for business leaders and professionals.
For November, Xiaopeng Li, AI Business Lead at Microsoft and Oslo AI co-founder will share insights about the fundamentals of AI technology as well as the responsible approach for developing and applying AI solutions.
On 16th of June, we'll be hosting the second edition of First steps in AI, an event for all AI passionates that would like to study or continue studying artificial intelligence in one of the available academic programs from Iași.
For this event, similar with the 2019 edition, our guests from Faculty of Computer Science, Mathematics, Automatic Control and Computer Engineering and Faculty of Electronics, Telecommunications and Information Technology will present the academic programs and courses that cover specific #AI topics in the upcoming academic year.
In May, we'll explore how to create Named Entity Recognition Systems. We're starting with identifying specific classes of entities from a given text-based dataset and learn how to install and use the necessary Python tools. In this month's talk we will show how to create, in a step-by-step manner, the components needed for creating custom NER systems.
Andrei, Data Scientist at SenseTask, will share how to do all this by leveraging Python's powerful set of open-source NLP libraries.
April is about tech and healthcare, as we'll be sharing insights into patients data, medical image processing and what are the best metrics and specific biases for machine learning models in medicine. More, we'll go through some relevant particularities of the domain knowledge in medical AI and the approach for developing AI projects using public medical datasets.
Charles Delingpole, CEO of ComplyAdvantage will join us for a fireside chat on artificial intelligence and how can machine learning help wiping out financial crime.
Meet Louison Dumont, Founder of Bitproof which developed Harmony, a cross platform voice agent builder. Using proprietary NLU and data augmentation methods, Harmony allows its users to easily build rich voice-enabled interfaces.
In January, we'll be entering into IBM world, learning how machine learning can obtain and provide valuable insights on customer behavior and dive into a brief look in the future and the usage of quantum computing when solving maching learning algorithms. More, Cristina, and Mihai, our speakers for this meetup will showcase IBM Auto AI and IBM Watson Studio.
Creating and retaining talent is one of the challenges that Romania, among other countries is having when talking about the development and adoption of artificial intelligence and other emerging technologies.
In December, together with our guests, we'll address this challenge in a panel-format meetup and learn what would be the right approach for AI education.
Our brain is capable of amazing feats like very efficient continuous learning from multi-modal data streaming from our senses. It uses mostly unsupervised learning and can be very frugal with data needed for learning and can quickly adapt to distribution shifts. And all while it uses about the same energy as a 20W light bulb.
In November, we’ll learn how this is possible and Alex, our speaker will present some deep learning architectures that try to get closer to brain capabilities using #neuroscience inspired ideas.
The use of face, as biometrics for authentication is becoming prevalent especially on mobile and embedded system. Alexandra, our speaker for this session will share how to overcome the challenges of constructing a robust verification system for devices with limited computation capabilities.
Starting August, besides our regular meetups focused on awareness and growth, we'll start to showcase startups or initiatives that are successfully applying AI and learn about their accomplishments and challenges.
For this month, our first invited startup is 𝐄𝐦𝐚𝐢𝐥𝐓𝐫𝐞𝐞 𝐀𝐈, a romanian startup which is using machine learning and NLP techniques to reinvent customer experience. #EmailTree AI comes as an end-to-end solution to increase email productivity, automate email composition and trigger automatic actions in databases.
For June, Petru Rebeja, Technical Lead at Centric IT Solutions will share insights on Variational Autoencoders, from both probabilistic and deep learning points of view to see why they aren't a generative model just like the others in the bunch.
When dealing with #NLP projects that are targeting Romanian language, we usually have difficulties finding good data and most of the times need to improvise.
In this talk, Andrei Pruteanu, Data Scientist and Content Author at PluralSight will describe why sentiment analysis is important for the Romanian language and the steps he did to obtain data and what NLP-specific tools used to create a sentiment analysis tool.
The topic for April will be #MedTech, where we’ll learn about algorithms that can enable the detection and prediction of epileptic seizures based on the activity of the brain, using CNN and RNN and take a sneak peek on how old-fashioned medical devices look like and state-of-the-art bionic arms and exoskeletons.
Our speaker, Erich Zimmermann, Computer Science student and Research Intern at the University of Southampton will share insights on the integration of AI in the use of medical technologies and how algorithms and available frameworks will enhance all medical devices and help our society towards a better and safer future.
In 2018, Deepmind unveiled Alphastar, one of most advanced AI capable to defeat top professional players in Starcraft.
Starcraft real-time strategy game was long time considered a grand challenge for #AI due to large action space, imperfect state information and no single superior strategy available. The game requires to constantly balance the short and long-term goals and adapt to unexpected situations.
Rudradeb Mitra started his career as an AI researcher in 2002 and worked with research labs like IBM Research Lab, TZI and startups and banks to build AI products. During that period he also published ten research papers on AI.
In the last two years, besides working on building AI products, he is also busy with sharing his knowledge and experiences. He is a mentor of Google Launchpad, MIT Enterprise, and Founders Institute, an Entrepreneur in Residence at WeWork Labs, and a senior AI advisor of EFMA.
Starting 2020 strong with a talk on decrypting a machine learning model. For February, Oriana Oniciuc, Data Scientist at E.ON will talk about one of the big challenges of Machine Learning - model explainability and share insights on how the research started in this field and how important is it to trust your model.
We’ll go into details with LIME (Local Interpretable Model-agnostic Explanations) and analyze a practical example of using LIME on a classification problem.
December is special and because of that, we’ll continue the tradition which we started last year, with a panel with representatives from academic medium and talk about research in AI, do a reality check and identify steps we need to do in order to stimulate the interest in researching.
After a quick review of the history of the #NLP breakthroughs, we will discuss the most recent architectures which attempt to give #AI a deeper understanding of language and visual concepts.
More concrete, we will discuss variants of #LSTM-based and more recent attention-based models and how they fare on harder NLP benchmarks.
For our 24th meetup, we're hosting a fireside chat with Valentin Irimia, an experienced technologist with broad experience, from fundamental algorithms and compiler construction to large scale distributed systems, low latency trading and machine learning infrastructure.
Currently, Valentin works for Google Deep Mind, a research company that uses science and technology to develop Artificial General Intelligence, with the purpose of advancing science and solving some of the problems faced by humanity.
On October 17th, 2019, together with deeplearning.ai we'll be hosting the very first Pie & AI webinar meetup.
The event will kick off with a live video talk from Dr. Andrew Ng. He will share exciting AI trends and tips and tricks for how to build a career in AI. We'll follow with a keynote and a panel discussion from local guest speakers. They'll also be taking questions from the audience.
The single most important trait of applying Big Data Analytics is undoubtedly empowerment.
It empowers stakeholders to take an informed business decision, it empowers every day commuters with the cumulative knowledge gained from all autonomous vehicles and so on... However, as we all know, with big power comes big responsibility and as the area covered by Big Data Analytics expands onto sensitive subjects it sparks the debate over morality.
In order to understand how AI can be used for advanced problem solving, practitioners need to acquire core machine learning concepts and start developing specific skills.
On Thursday, June 20th 2019, together with professors from "Alexandru Ioan Cuza" University Iași and "Gheorghe Asachi" Technical University Iași we’ll present the academic programs and courses that cover specific AI topics in the upcoming academic year (2019 - 2020).
What if cameras didn’t record any footage and just show what a human sees?
On April 16th, together with Virgil, AI and blockchain expert within NTT DATA Romania, we will explore how such systems are built and privacy implications.
For our #19th meetup we'll be covering some important aspects on safety and learn about the development phases of complex safety critical systems.
Daniel, our speaker for this edition is an assistant professor at Faculty of Electrical Engineering, "Gheorghe Asachi" Technical University of Iași and System Safety Engineer in R&D projects at System & Technologies, Continental Automotive Romania, where he's focused on image processing, human awareness monitoring, human-system interaction and safety engineering.
Alex Movila, Senior Java Developer at Fortech Iași will present a basic understanding and recent progress in generative modeling, focusing on the success of GANs (generative adversarial networks) in image generation.
#GANs gives imagination to AI, making possible amazing applications in design, photography, arts, semi-supervised training with fewer labelled data.
Andrei Sucila, Senior Applied Scientist at Amazon Development Center Romania will present how we can combine multiple supervised learning techniques into a single stronger learner, ensembles. Focusing on Super Learner (van der Laan, 2007) and subsemble learning, we'll see how it integrates in the classical ML pipeline. Prepare for a surprise algorithm at the end.
For December, as the last meetup for 2018, we’ve planned something different, a discussion panel where together with representatives from academic medium, we’ll discuss about AI opportunities.
Petru Rebeja, AI passionate and PhD student at Computer Science Faculty, focused on NLP and NN fields will provide insights of Recurrent Neural Networks from both theoretical and practical views and how a small and simple architecture can achieve great results.
Beyond targeted ads, talking homes and self-driving cars, artificial intelligence is about to reshape the core of our civilisation. Let's take a brief look at AI's impact on humanity's morals and ethics.
Our speaker for this event will be Antonio Caciuc, a general purpose hacker, software developer and recovering business analyst.
Artificial Intelligence is one of the pillars of many future products in various industries. Automotive is one of the future opportunities offered by AI, trying to achieve the goal of "zero accidents".
The event is hosted by Continental Romania, IAȘI AI community and Gheorghe Asachi Technical University of Iași.
Join us on Wednesday, June 27th for a joint event with #PINmagazine with the occasion of IAȘI AI community 1 year celebration and the launch of 9th issue of PINmagazine which has #artificial #intelligence as cover story.
Join IAȘI AI 11th meetup where, together with Ana Maria Simionovici we'll discuss about load prediction and balancing for cloud-based VoIP.
Daniel Voinea, the speaker for our 10th meetup will talk about traditional forecasting techniques such as ARIMA or Gaussian Processes that work well at small scale (1 series) but require human HPO and cannot "learn" from similar series.
Using deep learning, we explore solving this problem at scale (millions of time series) while trying to retain some traditional forecasting features such as confidence intervals and adding new ones: cold start forecasting, clustering via embeddings and low latency forecasting.
Cristian Padurariu, our speaker for IAȘI AI 9th meetup will discuss on how we can extract useful features from texts that can feed into machine learning algorithms. Questions will get answered during the meetup: What pre-processing of the text should we do? (stop words removal, word stemming etc.); What are the most common methods to transform words into vectors? (bag of words, tf-idf etc.); Where can we use this?
IAȘI AI kicks off 2018 with a new meetup where we'll showcase one of the AI initiatives started in Iași.
Andrei Iulian Marin (Founder of IRIS Robotics) and Silviu Filip (Software Engineer) will talk about humanoid robots, personal assistants with human face.
IAȘI AI welcomes you to its 7th meetup, where we'll discuss about data analysing and prediction.
Analyzing Data with Azure Machine Learning Studio and Python by Vlad Iliescu, Technical Director at Maxcode
History of AI has to include a chronology of dreams, hypes, failures and successes. Only by knowing how the field has evolved through the years, we can make educated guesses on what opportunities exist and what marketing traps to avoid.
Gabriel Marchidan, Software Architect at Feel IT Services will present a general outlook of hyperparameter optimization techniques, ranging from random search and grid search to CMA-ES (Co-variance Matrix Adaptation Evolution Strategy).
Emanuel Peia, Senior Developer at Endava will be our fourth speaker, presenting a general outline of human and machine learning pertaining to questions, exploration, attitudes and depth.
Ciprian Talmacel, Software Engineer at RomSoft will guide us into ConvNet (Convolutional Neural Networks) field and describe possible use cases as object classification, image segmentation, face recognition and more.
Andrei Șuiu, Software Architect at Ness Technologies, will be our speaker, sharing a personal view on AI applications in real world. He will also cover topics like modern AI projects, list of currently most significant public AI projects and their aims.
Being the first meetup of IAȘI AI community, the focus will be on meeting the AI enthusiasts of Iași and get to know each other.
More, we'll discuss about the goals we're trying to accomplish and next steps on how we can grow overall knowledge in this area.