Data Science Blog

MARCH 14, 2023

GPT-3 wurde mit mehr als 100 Milliarden Wörter trainiert, das parametrisierte Machine Learning Modell selbst wiegt 800 GB (quasi nur die Neuronen!) Neben Supervised Learning kam auch Reinforcement Learning zum Einsatz. April 2014 im Internet Archive ) auf: strata.oreilly.com. ChatGPT basiert auf GPT-3.5

Big Data 147

Big Data Big Data Apache Hadoop Data Science 147

Flipboard

FEBRUARY 28, 2023

The majority of companies developing the application-layer AI that’s driving the widespread adoption of the technology still rely on supervised learning, using large swaths of labeled training data. Currently, only well-funded institutions with access to a massive amount of GPU power are capable of building these models.

AI 132

AI AI ML ML 132

Explosion

MAY 17, 2023

In 2014 I started working on spaCy , and here’s an excerpt of how I explained the motivation for the library: Computers don’t understand text. Supervised learning is very strong for tasks such as text classification, entity recognition and relation extraction. That’s not a path to improvement. You need to be systematic.

Supervised Learning 135

Supervised Learning Natural Language Processing Clustering Machine Learning 135

Becoming Human

MARCH 16, 2023

Overhyped or not, investments in AI drug discovery jumped from $450 million in 2014 to a whopping $58 billion in 2021. AI drug discovery is exploding. Like in the human brain, these neurons work together to process information and make predictions or decisions.

AI 139

AI AI Deep Learning Deep Learning 139

Mlearning.ai

JULY 31, 2023

2014; Bojanowski et al., Data2Vec: A General Framework For Self-Supervised Learning in Speech, Vision and Language. Instead, why not use a set of embeddings that are already trained? Sometimes, this can be easier and much faster. Patch Embeddings What about images and audio files? References Baevski, A., and Auli, M.,

Deep Learning 52

Deep Learning Deep Learning Supervised Learning Algorithm 52

Google Research AI blog

JANUARY 18, 2023

There are a wide variety of approaches for generative models, which must learn to model complex data sets (e.g., Generative adversarial networks , developed in 2014, set up two models working against each other. natural images). Advances in generative image model capabilities over the past decade. Left: From I. Goodfellow, et al.

ML 132

ML ML AI AI 132

AssemblyAI

SEPTEMBER 12, 2023

Though once the industry standard, accuracy of these classical models had plateaued in recent years, opening the door for new approaches powered by advanced Deep Learning technology that’s also been behind the progress in other fields such as self-driving cars. End-to-end Deep Learning models are data hungry.

Deep Learning 52

Deep Learning Deep Learning Machine Learning Machine Learning 52