AI Flashcards
(32 cards)
encompasses a broad range of technologies that allow computers and
machines to perform tasks that typically require human intelligence.
Artificial Intelligence (AI)
AI is subdivided into 3 areas
Machine Learning (ML):
Deep Learning (DL):
Natural Language Processing (NLP)
Systems that learn from data without explicit programming.
Machine Learning (ML):
A subfield of ML using neural networks with many layers, capable of high
level abstractions.
Deep Learning (DL):
The interaction between computers and human (natural)
languages
Natural Language Processing (NLP):
AI methods, particularly __ and __ are ideally suited to
detect patterns, classify data, and make predictions that are often beyond human analytical capacity.
machine learning and deep learning,
____ such as CNNs and RNNs can predict splicing patterns, alternative transcripts, and
gene expression levels under different conditions.
Deep learning models
by Google uses a convolutional neural network to interpret sequencing data and
call genetic variants more accurately than traditional statistical methods.
DeepVariant
AI models use in proteomics (3)
-Predicting secondary, tertiary, and quaternary structures,
-Modeling protein-protein interactions,
-Predicting protein stability and folding dynamics.
The breakthrough of ___represents a massive leap, solving the decades-old problem of protein
folding by predicting structures directly from amino acid sequences with remarkable accuracy.
AlphaFold2
achieved protein structure predictions comparable to experimental results like X
ray crystallography and cryo-EM.
: AlphaFold2
are now capable of generating entirely new molecular structures (de novo drug
design) optimized for desired properties.
Deep learning models
uses deep learning to predict bioactivity of molecules, focusing on their ability to bind
to biological targets.
AtomNet
aims to understand complex biological systems as a whole, rather than in parts.
Systems biology
reconstruct pathways by learning from known interactions and suggesting new,
previously uncharacterized links.
Machine learning models
) are the backbone for analyzing images, extracting features, and
making accurate diagnostic predictions.
Convolutional neural networks (CNNs
Breakthrough in predicting protein 3D structures.
Protein
Prediction
CNN-based caller outperforming classical statistical
methods.
Genomic
Detection
Deep learning to predict binding affinity for drug
design.
Drug Discovery
Automatic cancer diagnosis from tissue images.
Cancer Detection
Design
Automatic cancer diagnosis from tissue images.
Optimizing gene circuits and metabolic pathways using
ML.
Synthetic Biology
___models make predictions without explaining how decisions are made.
Black-box AI
AI is now being used to analyze multiple layers of biological data
simultaneously to get a comprehensive view of biological processes.
Integrative Omics with AI:
Single-cell RNA-seq and other assays generate sparse but highly informative
data. AI is crucial for analyzing this complexity.
Single-Cell Biology:
