The semantic matching capabilities of neural information retrieval can ameliorate synonymy and polysemy problems of symbolic approaches. However, neural models' dense representations are more suitable for re-ranking, due to their inefficiency. Sparse
representations, either in symbolic or latent form, are more efficient with an inverted index. Taking the merits of the sparse and dense representations, we propose an ultra-high dimensional (UHD) representation scheme equipped with directly controllable sparsity. UHD's large capacity and minimal noise and interference among the dimensions allow for binarized representations, which are highly efficient for storage and search. Also proposed is a bucketing method, where the embeddings from multiple layers of BERT are selected/merged to represent diverse linguistic aspects. We test our models with MS MARCO and TREC CAR, showing that our models outperforms other sparse models.
A text retrieval system for language learning returns reading materials at the appropriate difficulty level for the user. The system typically maintains a learner model on the user's vocabulary knowledge, and identifies texts that best fit the model.
As the user's language proficiency increases, model updates are necessary to retrieve texts with the corresponding lexical complexity. We investigate an open learner model that allows user modification of its content, and evaluate its effectiveness with respect to the amount of user update effort. We compare this model with the graded approach, in which the system returns texts at the optimal grade. When the user makes at least half of the expected updates to the open learner model, simulation results show that it outperforms the graded approach in retrieving texts that fit user preference for new-word density.
Abstract Dual encoders perform retrieval by encoding documents and queries into dense low-dimensional vectors, scoring each document by its inner product with the query. We investigate the capacity of this architecture relative to sparse bag-of-words
models and attentional neural networks. Using both theoretical and empirical analysis, we establish connections between the encoding dimension, the margin between gold and lower-ranked documents, and the document length, suggesting limitations in the capacity of fixed-length encodings to support precise retrieval of long documents. Building on these insights, we propose a simple neural model that combines the efficiency of dual encoders with some of the expressiveness of more costly attentional architectures, and explore sparse-dense hybrids to capitalize on the precision of sparse retrieval. These models outperform strong alternatives in large-scale retrieval.
