61 citations · 115 across the 19 of their papers we have counts for
20 papers · 1 filter
ImpRAG: Retrieval-Augmented Generation with Implicit Queries
Wenzheng Zhang, Xi Victoria Lin, Karl Stratos +2
Retrieval-Augmented Generation (RAG) systems traditionally treat retrieval and generation as separate processes, requiring explicit textual queries to connect them. This separation…
Rethinking Reflection in Pre-Training
Essential AI, :, Darsh J Shah +26
A language model's ability to reflect on its own reasoning provides a key advantage for solving complex problems. While most recent research has focused on how this ability develop…
Seq2seq is All You Need for Coreference Resolution
Wenzheng Zhang, Sam Wiseman, Karl Stratos
Existing works on coreference resolution suggest that task-specific models are necessary to achieve state-of-the-art performance. In this work, we present compelling evidence that…
Improving Multitask Retrieval by Promoting Task Specialization
Wenzheng Zhang, Chenyan Xiong, Karl Stratos +1
In multitask retrieval, a single retriever is trained to retrieve relevant contexts for multiple tasks. Despite its practical appeal, naive multitask retrieval lags behind task-spe…
EntQA: Entity Linking as Question Answering
Wenzheng Zhang, Wenyue Hua, Karl Stratos
A conventional approach to entity linking is to first find mentions in a given document and then infer their underlying entities in the knowledge base. A well-known limitation of t…
Understanding Hard Negatives in Noise Contrastive Estimation
Wenzheng Zhang, Karl Stratos
The choice of negative examples is important in noise contrastive estimation. Recent works find that hard negatives -- highest-scoring incorrect examples under the model -- are eff…
Fast and Effective Biomedical Entity Linking Using a Dual Encoder
Rajarshi Bhowmik, Karl Stratos, Gerard de Melo
Biomedical entity linking is the task of identifying mentions of biomedical concepts in text documents and mapping them to canonical entities in a target thesaurus. Recent advancem…
Data-to-text Generation by Splicing Together Nearest Neighbors
Sam Wiseman, Arturs Backurs, Karl Stratos
We propose to tackle data-to-text generation tasks by directly splicing together retrieved segments of text from "neighbor" source-target pairs. Unlike recent work that conditions…
Unsupervised Label Refinement Improves Dataless Text Classification
Zewei Chu, Karl Stratos, Kevin Gimpel
Dataless text classification is capable of classifying documents into previously unseen labels by assigning a score to any document paired with a label description. While promising…
Corrected CBOW Performs as well as Skip-gram
Ozan İrsoy, Adrian Benton, Karl Stratos
Mikolov et al. (2013a) observed that continuous bag-of-words (CBOW) word embeddings tend to underperform Skip-gram (SG) embeddings, and this finding has been reported in subsequent…
Mining Knowledge for Natural Language Inference from Wikipedia Categories
Mingda Chen, Zewei Chu, Karl Stratos +1
Accurate lexical entailment (LE) and natural language inference (NLI) often require large quantities of costly annotations. To alleviate the need for labeled data, we introduce Wik…
NatCat: Weakly Supervised Text Classification with Naturally Annotated Resources
Zewei Chu, Karl Stratos, Kevin Gimpel
We describe NatCat, a large-scale resource for text classification constructed from three data sources: Wikipedia, Stack Exchange, and Reddit. NatCat consists of document-category…
Discrete Latent Variable Representations for Low-Resource Text Classification
Shuning Jin, Sam Wiseman, Karl Stratos +1
While much work on deep latent variable models of text uses continuous latent variables, discrete latent variables are interesting because they are more interpretable and typically…
EntEval: A Holistic Evaluation Benchmark for Entity Representations
Mingda Chen, Zewei Chu, Yang Chen +2
Rich entity representations are useful for a wide class of problems involving entities. Despite their importance, there is no standardized benchmark that evaluates the overall qual…
Label-Agnostic Sequence Labeling by Copying Nearest Neighbors
Sam Wiseman, Karl Stratos
Retrieve-and-edit based approaches to structured prediction, where structures associated with retrieved neighbors are edited to form new structures, have recently attracted increas…
Mutual Information Maximization for Simple and Accurate Part-Of-Speech Induction
Karl Stratos
We address part-of-speech (POS) induction by maximizing the mutual information between the induced label and its context. We focus on two training objectives that are amenable to s…
OneNet: Joint Domain, Intent, Slot Prediction for Spoken Language Understanding
Young-Bum Kim, Sungjin Lee, Karl Stratos
In practice, most spoken language understanding systems process user input in a pipelined manner; first domain is predicted, then intent and semantic slots are inferred according t…
Reconstruction of Word Embeddings from Sub-Word Parameters
Karl Stratos
Pre-trained word embeddings improve the performance of a neural model at the cost of increasing the model size. We propose to benefit from this resource without paying the cost by…
A Sub-Character Architecture for Korean Language Processing
Karl Stratos
We introduce a novel sub-character architecture that exploits a unique compositional structure of the Korean language. Our method decomposes each character into a small set of prim…
Entity Identification as Multitasking
Karl Stratos
Standard approaches in entity identification hard-code boundary detection and type prediction into labels (e.g., John/B-PER Smith/I-PER) and then perform Viterbi. This has two disa…