Sökning: "Dynamic Embeddings"

Visar resultat 1 - 5 av 11 uppsatser innehållade orden Dynamic Embeddings.

1. 1. Learning the shapes of protein pockets

   Master-uppsats, Göteborgs universitet/Institutionen för data- och informationsteknik

   Författare :Alejandro Corrochano; Yossra Gharbi; [2022-10-14]
   Nyckelord :Protein; cavity; ligand-binding; 3D-equivariance; shape; latent space; e3nn; Fpocket; sc-PDB;

   Sammanfattning : The comparison of protein pockets plays an important role in drug discovery. Through the identification of binding sites with similar structures, we can assist in finding hits and characterizing the function of proteins. Traditionally, the geometry of cavities has been described with scalar features, which are not fully representative of the shape. LÄS MER

3. 2. Reliable graph predictions : Conformal prediction for Graph Neural Networks

   Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

   Författare :Albin Bååw; [2022]
   Nyckelord :Conformal prediction; Graph Neural Networks; Dynamic graphs; Distribution shift; Coverage gap; Konform prediktion; Neurala Nätverk för Grafer; Dynamiska grafer; Distributionsförändring; täckningsgap;

   Sammanfattning : We have seen a rapid increase in the development of deep learning algorithms in recent decades. However, while these algorithms have unlocked new business areas and led to great development in many fields, they are usually limited to Euclidean data. LÄS MER

4. 3. LSTM Feature Engineering Through Time Series Similarity Embedding

   Master-uppsats, Linköpings universitet/Institutionen för datavetenskap

   Författare :Sebastian Bångerius; [2022]
   Nyckelord :Embedding; time series; LSTM; feature engineering; DTW; correlation; prediction;

   Sammanfattning : Time series prediction has many applications. In cases with simultaneous series (like measurements of weather from multiple stations, or multiple stocks on the stock market)it is not unlikely that these series from different measurement origins behave similarly, or respond to the same contextual signals. LÄS MER

5. 4. Comparison of state-of-the-art Temporal Interaction Network methods in different settings : Novel models to predict temporal behavior

   Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

   Författare :Indre Tauroseviciute; [2021]
   Nyckelord :Recommendation systems; Neural Collaborative Filtering; RNN; Backpropagation; Comparative analysis;

   Sammanfattning : Recommendation systems become more and more necessary due to the growing supply chain. Therefore, scientists are developing models that can serve different recommendation needs faster than before, and it is getting more complicated to choose the model for a specific case. LÄS MER

7. 5. Dynamic Graph Representation Learning on Enterprise Live Video Streaming Events

   Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

   Författare :Achilleas Stefanidis; [2020]
   Nyckelord :Dynamic graph representation learning; Selfattention mechanism; Dynamic link prediction; Inlärning av dynamisk grafrepresentation; Koncentrationsmekanism; Dynamisk länkprognos;

   Sammanfattning : Enterprises use live video streaming as a mean of communication. Streaming high-quality video to thousands of devices in a corporate network is not an easy task; the bandwidth requirements often exceed the network capacity. LÄS MER