Abstract: The technology provides neural network models having dual encoder architectures. For instance, a token embedder layer section of a dual encoder is associated with a first input and a second input and generates token embeddings (2002). An encoder layer section receives the token embeddings from the token embedder layer and generates encodings based on the token embeddings (2004). A projection layer receives the encodings from the encoder section and generates a set of projections (2006). The projection layer is shared by the dual encoders (FIG. 4C). An embedding space generates, based on the set of projections, a question embedding and a response embedding (2008). These question and response embeddings are used in identifying a set of candidate responses to an input question (2008). The embedding space is configured during training according to a contrastive loss that applies same tower negatives to generate the question and response embeddings (2010).

Abstract: According to one general aspect, a method is provided for managing memory when counting unique items, the method using a pattern of bits in a unique estimator mask. The method may create a unique estimator mask based on fingerprints calculated for previously encountered items, and determine a number with the highest probability for creating the pattern of bits in the mask. When the number with the highest probability is determined, it may be returned as the estimated count of unique items.

Abstract: According to one general aspect, a method is provided for managing memory when tracking an estimated number of events, the method using a logarithmic count to track the estimated number. The method may detect a new occurrence of the event and increment the logarithmic count when no bits in a random number are in bit positions of significance. The bit positions of significance may be based on the value of the logarithmic counter and may be represented by bits having a non-zero value in a bitmask.