Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an encoder neural network configured to receive a data item and to process the data item to output a compressed representation of the data item. In one aspect, a method includes, for each training data item: processing the data item using the encoder neural network to generate a latent representation of the training data item; processing the latent representation using a hyper-encoder neural network to determine a conditional entropy model; generating a compressed representation of the training data item; processing the compressed representation using a decoder neural network to generate a reconstruction of the training data item; processing the reconstruction of the training data item using a discriminator neural network to generate a discriminator network output; evaluating a first loss function; and determining an update to the current values of the encoder network parameters.

Abstract: A first image and textual content associated with the first image is obtained. A second image that depicts the textual content associated with the first image is rendered. The first image and the second image are processed with a machine-learned encoding model to respectively obtain a first image embedding and a second image embedding for an image embedding space including a plurality of image embeddings. The machine-learned encoding model is trained based on a difference between the first image embedding and the second image embedding.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a neural network that is configured to process an input image to generate a network output for the input image. In one aspect, a method comprises, at each of a plurality of training steps: obtaining a plurality of training images for the training step; obtaining, for each of the plurality of training images, a respective target output; and selecting, from a plurality of image patch generation schemes, an image patch generation scheme for the training step, wherein, given an input image, each of the plurality of image patch generation schemes generates a different number of patches of the input image, and wherein each patch comprises a respective subset of the pixels of the input image.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an encoder neural network configured to receive a data item and to process the data item to output a compressed representation of the data item. In one aspect, a method includes, for each training data item: processing the data item using the encoder neural network to generate a latent representation of the training data item; processing the latent representation using a hyper-encoder neural network to determine a conditional entropy model; generating a compressed representation of the training data item; processing the compressed representation using a decoder neural network to generate a reconstruction of the training data item; processing the reconstruction of the training data item using a discriminator neural network to generate a discriminator network output; evaluating a first loss function; and determining an update to the current values of the encoder network parameters.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an encoder neural network configured to receive a data item and to process the data item to output a compressed representation of the data item. In one aspect, a method includes, for each training data item: processing the data item using the encoder neural network to generate a latent representation of the training data item; processing the latent representation using a hyper-encoder neural network to determine a conditional entropy model; generating a compressed representation of the training data item; processing the compressed representation using a decoder neural network to generate a reconstruction of the training data item; processing the reconstruction of the training data item using a discriminator neural network to generate a discriminator network output; evaluating a first loss function; and determining an update to the current values of the encoder network parameters.

Abstract: The invention relates to the use of vinyl acetate-copolymers as a shrinkage-reducing additive (low profile additive) in cold-curing systems for producing composite materials, characterized in that vinyl acetate-copolyerisate of 40 to 95 wt % vinyl acetate and 5 to 60 wt % of one or more co-monomers from the group containing vinyl esters of unbranched or branched carboxylic acids having 3 to 20 C-atoms and methacrylic acid esters and acrylic acid esters of unbranched or branched alcohols having 2 to 15 C-atoms are used, wherein the specifications in wt % relate to the total weight of the co-monomers and add up to 100 wt %.

Abstract: A low profile additive includes vinyl acetate-isopropenyl acetate copolymers. The vinyl acetate-isopropenyl acetate copolymers are based on from 2 to 98% by weight of vinyl acetate, from 2 to 98% by weight of isopropenyl acetate and optionally one or more further ethylenically unsaturated monomers, in each case based on the total weight of the vinyl acetate-isopropenyl acetate copolymers.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an encoder neural network configured to receive a data item and to process the data item to output a compressed representation of the data item. In one aspect, a method includes, for each training data item: processing the data item using the encoder neural network to generate a latent representation of the training data item; processing the latent representation using a hyper-encoder neural network to determine a conditional entropy model; generating a compressed representation of the training data item; processing the compressed representation using a decoder neural network to generate a reconstruction of the training data item; processing the reconstruction of the training data item using a discriminator neural network to generate a discriminator network output; evaluating a first loss function; and determining an update to the current values of the encoder network parameters.

Abstract: The present invention relates to binding proteins specific for VEGF-A, in particular to recombinant binding proteins comprising a polyethylene glycol moiety and a binding domain, which inhibits VEGF-Axxx binding to VEGFR-2. Examples of such recombinant binding proteins are proteins which comprise an ankyrin repeat domain with the desired binding specificity, and a polyethylene glycol moiety. The binding proteins are useful in the treatment of cancer and other pathological conditions, e.g. eye diseases such as age-related macular degeneration.

Abstract: The invention relates to the use of vinyl acetate-copolymers as a shrinkage-reducing additive (low profile additive) in cold-curing systems for producing composite materials, characterized in that vinyl acetate-copolyerisate of 40 to 95 wt % vinyl acetate and 5 to 60 wt % of one or more co-monomers from the group containing vinyl esters of unbranched or branched carboxylic acids having 3 to 20 C-atoms and methacrylic acid esters and acrylic acid esters of unbranched or branched alcohols having 2 to 15 C-atoms are used, wherein the specifications in wt % relate to the total weight of the co-monomers and add up to 100 wt %.

Abstract: The present invention relates to binding proteins specific for VEGF-A, in particular to recombinant binding proteins comprising a polyethylene glycol moiety and a binding domain, which inhibits VEGF-Axxx binding to VEGFR-2. Examples of such recombinant binding proteins are proteins which comprise an ankyrin repeat domain with the desired binding specificity, and a polyethylene glycol moiety. The binding proteins are useful in the treatment of cancer and other pathological conditions, e.g. eye diseases such as age-related macular degeneration.

Abstract: The present invention relates to binding proteins specific for VEGF-A, in particular to recombinant binding proteins comprising a polyethylene glycol moiety and a binding domain, which inhibits VEGF-Axxx binding to VEGFR-2. Examples of such recombinant binding proteins are proteins which comprise an ankyrin repeat domain with the desired binding specificity, and a polyethylene glycol moiety. The binding proteins are useful in the treatment of cancer and other pathological conditions, e.g. eye diseases such as age-related macular degeneration.

Abstract: The present invention relates to binding proteins specific for VEGF-A, in particular to recombinant binding proteins comprising a polyethylene glycol moiety and a binding domain, which inhibits VEGF-Axxx binding to VEGFR-2. Examples of such recombinant binding proteins are proteins which comprise an ankyrin repeat domain with the desired binding specificity, and a polyethylene glycol moiety. The binding proteins are useful in the treatment of cancer and other pathological conditions, e.g. eye diseases such as age-related macular degeneration.

Abstract: A repeat protein from a collection of repeat proteins, wherein each repeat protein of said collection comprises a repeat domain, which comprises a set of consecutive repeat modules, wherein the repeat modules have the same fold and stack tightly to create a superhelical structure having a joint hydrophobic core, wherein each of the repeat modules is derived from one or more repeat units and wherein the repeat units comprise framework residues, which contribute to the folding topology of the repeat unit or contribute to an interaction with a neighboring repeat unit, and target interaction residues, which contribute to an interaction with a target substance, wherein the repeat proteins of the collection differ from other repeat proteins in the collection in at least one amino acid position of the repeat modules is described as are related pharmaceuticals and nucleic acid molecules.

Abstract: The present invention relates to binding proteins specific for VEGF-A, in particular to recombinant binding proteins comprising a polyethylene glycol moiety and a binding domain, which inhibits VEGF-Axxx binding to VEGFR-2. Examples of such recombinant binding proteins are proteins which comprise an ankyrin repeat domain with the desired binding specificity, and a polyethylene glycol moiety. The binding proteins are useful in the treatment of cancer and other pathological conditions, e.g. eye diseases such as age-related macular degeneration.

Abstract: The invention relates to collections of target-specific designed binding proteins based on armadillo repeat proteins, and to a method of generating them. Designed armadillo repeat proteins are based on consensus sequences of single armadillo repeat units. These repeat proteins can be used as scaffolds for peptide recognition. Such a scaffold provides a binding mode that is in principle the same for every small recognizable unit, e.g. an amino acid or dipeptide, allowing a precise and modular recognition of a peptide in extended conformation. The method allows to generate a series of modules recognizing these simple units, and to combine such building blocks to create a binding site for any desired peptide target without performing additional selections.

Abstract: The present invention relates to binding proteins specific for Vascular Endothelial Growth Factor A (VEGF-A), in particular to recombinant binding proteins comprising a polyethylene glycol moiety and a binding domain, which inhibits VEGF-Axxx (wherein xxx denotes the amino acid length of the VEGF-A mature protein) binding to Vascular Endothelial Growth Factor Receptor 2 (VEGFR-2). Examples of such recombinant binding proteins are proteins which comprise an ankyrin repeat domain with the desired binding specificity, and a polyethylene glycol moiety. The binding proteins are useful in the treatment of cancer and other pathological conditions, e.g. eye diseases such as age-related macular degeneration.

Abstract: The present invention relates to binding proteins specific for VEGF-A, in particular to recombinant binding proteins comprising a binding domain, which inhibits VEGF-Axxx binding to VEGFR-2. Examples of such binding proteins are proteins which comprise an ankyrin repeat domain with the desired binding specificity. The binding proteins are useful in the treatment of cancer and other pathological conditions, e.g. eye diseases such as age-related macular degeneration.

Abstract: A repeat protein from a collection of repeat proteins, wherein each repeat protein of said collection comprises a repeat domain, which comprises a set of consecutive repeat modules, wherein the repeat modules have the same fold and stack tightly to create a superhelical structure having a joint hydrophobic core, wherein each of the repeat modules is derived from one or more repeat units and wherein the repeat units comprise framework residues, which contribute to the folding topology of the repeat unit or contribute to an interaction with a neighboring repeat unit, and target interaction residues, which contribute to an interaction with a target substance, wherein the repeat proteins of the collection differ from other repeat proteins in the collection in at least one amino acid position of the repeat modules is described as are related pharmaceuticals and nucleic acid molecules.

Abstract: The invention relates to the use of polymers on the basis of renewable resources as LPA, characterized in that the polymers on the basis of renewable resources can be obtained by the radically initiated polymerization of one or more ethylenically unsaturated monomers in the presence of one or more renewable resources selected from the group consisting of carbohydrates and natural rubbers.