More Like this

1. A Hybrid Deep Model for Learning to Rank Data Tables

   https://doi.org/10.1109/BigData50022.2020.9378185  

   Trabelsi, Mohamed; Chen, Zhiyu; Davison, Brian D.; Heflin, Jeff (December 2020, 2020 IEEE International Conference on Big Data (Big Data))

   null (Ed.)

   We address the problem of ad hoc table retrieval via a new neural architecture that incorporates both semantic and relevance matching. Understanding the connection between the structured form of a table and query tokens is an important yet neglected problem in information retrieval. We use a learning- to-rank approach to train a system to capture semantic and relevance signals within interactions between the structured form of candidate tables and query tokens. Convolutional filters that extract contextual features from query/table interactions are combined with a feature vector based on the distributions of term similarity between queries and tables. We propose using row and column summaries to incorporate table content into our new neural model. We evaluate our approach using two datasets, and we demonstrate substantial improvements in terms of retrieval metrics over state-of-the-art methods in table retrieval and document retrieval, and neural architectures from sentence, document, and table type classification adapted to the table retrieval task. Our ablation study supports the importance of both semantic and relevance matching in the table retrieval. 

   more » « less
2. Investigating Adaptation to And-Coordination in English: An ERP Study

   https://doi.org/10.3390/languages8030207  

   Kaan, Edith (September 2023, Languages)

   According to certain approaches to adaptation, readers and listeners quickly adjust their processing of sentences to match properties of recently encountered sentences. The present preregistered study used ERP (event-related brain potentials) to investigate how and when readers change their processing in response to recent exposure to sentences of a particular structure. We presented English speakers (n = 36) with three virtual blocks of English sentences with and-coordination ambiguities. In the first and third block, the ambiguity was always resolved towards a noun phrase (NP-) coordination; in the second block, the structure was always a clausal (S-) coordination. We manipulated the plausibility of the critical noun after the conjunct. N400 and P600 plausibility effects were probed to see to what extent the reader preferred an NP- coordination or expected the sentence to continue differently. Our results suggest that readers change their processing as a function of recent exposure but that they do not immediately adapt to the target structure. Furthermore, we observed substantial individual variation in the type and change in response over the course of the study. The idea that structural adaptation is immediate and a direct reflection of the properties of the recent context therefore needs to be fine-tuned. 

   more » « less
3. Unsupervised Partial Sentence Matching for Cited Text Identification

   Ricci, Kathryn; Chang, Haw-Shiuan; Goyal, Purujit; McCallum, Andrew (October 2022, ACL Proceedings of the Third Workshop on Scholarly Document Processing)

   Given a citation in the body of a research paper, cited text identification aims to find the sentences in the cited paper that are most relevant to the citing sentence. The task is fundamentally one of sentence matching, where affinity is often assessed by a cosine similarity between sentence embeddings. However, (a) sentences may not be well-represented by a single embedding because they contain multiple distinct semantic aspects, and (b) good matches may not require a strong match in all aspects. To overcome these limitations, we propose a simple and efficient unsupervised method for cited text identification that adapts an asymmetric similarity measure to allow partial matches of multiple aspects in both sentences. On the CL-SciSumm dataset we find that our method outperforms a baseline symmetric approach, and, surprisingly, also outperforms all supervised and unsupervised systems submitted to past editions of CL-SciSumm Shared Task 1a. 

   more » « less
4. Unsupervised Discovery of Multimodal Links in Multi-image, Multi-sentence Documents

   https://doi.org/10.18653/v1/D19-1210  

   Hessel, Jack; Lee, Lillian; Mimno, David (November 2019, Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP))

   null (Ed.)

   Images and text co-occur constantly on the web, but explicit links between images and sentences (or other intra-document textual units) are often not present. We present algorithms that discover image-sentence relationships without relying on explicit multimodal annotation in training. We experiment on seven datasets of varying difficulty, ranging from documents consisting of groups of images captioned post hoc by crowdworkers to naturally-occurring user-generated multimodal documents. We find that a structured training objective based on identifying whether collections of images and sentences co-occur in documents can suffice to predict links between specific sentences and specific images within the same document at test time. 

   more » « less
5. Bridging Continuous and Discrete Spaces: Interpretable Sentence Representation Learning via Compositional Operations

   https://doi.org/10.18653/v1/2023.emnlp-main.900  

   Huang, James; Yao, Wenlin; Song, Kaiqiang; Zhang, Hongming; Chen, Muhao; Yu, Dong (January 2023, Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing)

   Traditional sentence embedding models encode sentences into vector representations to capture useful properties such as the semantic similarity between sentences. However, in addition to similarity, sentence semantics can also be interpreted via compositional operations such as sentence fusion or difference. It is unclear whether the compositional semantics of sentences can be directly reflected as compositional operations in the embedding space. To more effectively bridge the continuous embedding and discrete text spaces, we explore the plausibility of incorporating various compositional properties into the sentence embedding space that allows us to interpret embedding transformations as compositional sentence operations. We propose InterSent, an end-to-end framework for learning interpretable sentence embeddings that supports compositional sentence operations in the embedding space. Our method optimizes operator networks and a bottleneck encoder-decoder model to produce meaningful and interpretable sentence embeddings. Experimental results demonstrate that our method significantly improves the interpretability of sentence embeddings on four textual generation tasks over existing approaches while maintaining strong performance on traditional semantic similarity tasks. 

   more » « less