| Challenge: | Functional Distributional Semantics models the meaning of a word as a binary classifier rather than a numerical vector. |
| Approach: | They propose a method to train a Functional Distributional Semantics model with grounded visual data. |
| Outcome: | The proposed model outperforms previous work on learning semantics from Visual Genome on four external evaluation datasets. |
Similar Papers
Autoencoding Pixies: Amortised Variational Inference with Graph Convolutions for Functional Distributional Semantics (2020.acl-main)
Copied to clipboard
| Challenge: | Functional Distributional Semantics represents the meaning of a word as a function (a binary classifier), instead of . a vector. |
| Approach: | They propose a framework which augments the generative model of Functional Distributional Semantics with a graph-convolutional neural network to perform amortised variational inference. |
| Outcome: | The proposed framework outperforms BERT, a large pre-trained language model on two tasks and outperformed other approaches. |
Semantic Specialization of Distributional Word Vectors (D19-2)
Copied to clipboard
| Challenge: | Distributional word vectors conflate various paradigmatic and syntagmatic lexico-semantic relations. |
| Approach: | This tutorial provides an overview of specialization methods for distributional word vectors . a common solution is to include external lexico-semantic knowledge in a reshaped vector space . |
| Outcome: | This paper provides an overview of specialization methods for distributional word vectors . the most recent developments include a new method for asymmetric relations in Euclidean . |
What are the Goals of Distributional Semantics? (2020.acl-main)
Copied to clipboard
| Challenge: | a new paper examines distributional semantic models' ability to deal with semantic challenges . authors argue that assessing progress in any field requires explicit long-term goals . |
| Approach: | They propose a broad linguistic perspective to assess distributional semantic models' ability to deal with various semantic challenges. |
| Outcome: | The proposed models can handle various semantic challenges, but they need to be explicit . a top-down approach is largely bottom-up, while a bottom-down one is mainly top-up . the authors argue that the goal is unclear and that the models are not scalable . |
Beyond Facts- Benchmarking Distributional Reading Comprehension in Large Language Models (2026.findings-acl)
Copied to clipboard
Pei-Fu Guo, Ya An Tsai, Chun-Chia Hsu, Kai-Xin Chen, Yun-Da Tsai, Kai-Wei Chang, Nanyun Peng, Mi-Yen Yeh, Shou-De Lin
| Challenge: | Existing reading comprehension benchmarks focus on factual information, but many real-world tasks require distributional knowledge expressed across text. |
| Approach: | They propose a reading comprehension benchmark for LLMs to evaluate their ability to infer distributional knowledge from natural language. |
| Outcome: | Experiments with multiple LLMs show that the model outperforms baselines, but performance varies widely across distribution types and characteristics. |
Function Words as Statistical Cues for Language Learning (2026.acl-long)
Copied to clipboard
| Challenge: | Existing studies have argued that function words aid learning abstract grammatical knowledge from linear input. |
| Approach: | They examine the statistical distribution of function words and their properties . they show that function words are reliable, diverse, and informative . |
| Outcome: | The results show that function words preserve high frequency, reliable syntactic association, phrase-boundary alignment and are informative to structural dependency. |
A Distributional Perspective on Word Learning in Neural Language Models (2025.naacl-long)
Copied to clipboard
| Challenge: | Language models are increasingly being studied as models of human language learners. |
| Approach: | They propose a distributional approach to word learning that captures distributional knowledge and gradient preferences for the word’s appropriateness. |
| Outcome: | The proposed signatures capture knowledge of where the target word can and cannot occur as well as gradient preferences about the word’s appropriateness. |
Learning Visually-Grounded Semantics from Contrastive Adversarial Samples (C18-1)
Copied to clipboard
| Challenge: | Existing frameworks for grounding distributional representations of texts on the visual domain are limited . effective and efficient grounding of distributional embeddings remains challenging . |
| Approach: | They propose to ground distributional representations of texts on the visual domain using visual-semantic embeddings. |
| Outcome: | The proposed model improves on a diverse set of downstream tasks and defends known-type adversarial attacks. |
Representing Verbs with Visual Argument Vectors (2020.lrec-1)
Copied to clipboard
| Challenge: | Existing models for verb semantic similarities are based on linguistic data, but they do not register intuitive attributes. |
| Approach: | They evaluated two textual distributional semantic models and a visual one to explore verb semantic similarities. |
| Outcome: | The proposed models extract meaningful information and capture semantic similarity between verbs using visual distributional models. |
Short-Term Meaning Shift: A Distributional Exploration (N19-1)
Copied to clipboard
| Challenge: | a new study examines the phenomenon of short-term meaning shift in online communities . the authors use distributional representations to explore the phenomenon . |
| Approach: | They propose to use distributional representations to explore short-term meaning shift in online communities. |
| Outcome: | The proposed model has problems distinguishing meaning shift from referential phenomena, and measures contextual variability to remedy this. |
Why is penguin more similar to polar bear than to sea gull? Analyzing conceptual knowledge in distributional models (2020.acl-srw)
Copied to clipboard
| Challenge: | Several analysis methods have been shown to be limited and are not well understood . thesis aims to understand distributional semantic representations based on linguistic data . |
| Approach: | They propose a framework for investigating the information encoded in distributional semantic models . they combine observations made on corpora with insights obtained from data manipulation experiments . |
| Outcome: | The proposed framework pairs observations made on corpora with insights obtained from data manipulation experiments. |