Open AccessPosted Content
Can Language Models Encode Perceptual Structure Without Grounding? A Case Study in Color
TL;DR: The authors found that warmer colors are, on average, better aligned to the perceptual color space than cooler ones, suggesting an intriguing connection to findings from recent work on efficient communication in color naming.
read more
Abstract: Pretrained language models have been shown to encode relational information, such as the relations between entities or concepts in knowledge-bases -- (Paris, Capital, France). However, simple relations of this type can often be recovered heuristically and the extent to which models implicitly reflect topological structure that is grounded in world, such as perceptual structure, is unknown. To explore this question, we conduct a thorough case study on color. Namely, we employ a dataset of monolexemic color terms and color chips represented in CIELAB, a color space with a perceptually meaningful distance metric.
Using two methods of evaluating the structural alignment of colors in this space with text-derived color term representations, we find significant correspondence. Analyzing the differences in alignment across the color spectrum, we find that warmer colors are, on average, better aligned to the perceptual color space than cooler ones, suggesting an intriguing connection to findings from recent work on efficient communication in color naming. Further analysis suggests that differences in alignment are, in part, mediated by collocationality and differences in syntactic usage, posing questions as to the relationship between color perception and usage and context.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Symbols and grounding in large language models
TL;DR: The authors argue that the answer depends on models' underlying competence, and thus that the focus of the debate should be on empirical work which seeks to characterize the representations and processing algorithms that underlie model behaviour.
53
Do Large Language Models Know What Humans Know?
TL;DR: This article found that the language model significantly exceeds chance behavior, but it does not explain the full extent of their behavior, despite being exposed to more language than a human would in a lifetime, while statistical learning from language exposure may in part explain how humans develop the ability to reason about the mental states of others.
52
Generating meaning: active inference and the scope and limits of passive AI
Giovanni Pezzulo,Thomas Parr,Paul Cisek,Andy Clark,Karl Friston +4 more
TL;DR: The generative models of living organisms are inextricably anchored to the body and world and must capture and control the sensory consequences of action, which allows embodied agents to intervene upon their worlds in ways that constantly put their best models to the test.
34
Generative Artificial Intelligence in Education: From Deceptive to Disruptive
Marc Alier Forment,Francisco J. García-Peñalvo,Jorge D. Camba +2 more
TL;DR: The need for implementing ethical practices in the use of GenAI models and ensuring that the technology is used to support and not replace the student's learning experience is highlighted.
31
Strong Prediction: Language model surprisal explains multiple N400 effects
TL;DR: The authors used a neural language model trained to compute the conditional probability of any word based on the words that precede it to operationalize contextual predictability, using an information theoretical construct known as surprisal.
References
•Posted Content
RoBERTa: A Robustly Optimized BERT Pretraining Approach
Yinhan Liu,Myle Ott,Naman Goyal,Jingfei Du,Mandar Joshi,Danqi Chen,Omer Levy,Michael Lewis,Luke Zettlemoyer,Veselin Stoyanov +9 more
TL;DR: It is found that BERT was significantly undertrained, and can match or exceed the performance of every model published after it, and the best model achieves state-of-the-art results on GLUE, RACE and SQuAD.
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Jacob Devlin,Ming-Wei Chang,Kenton Lee,Kristina Toutanova +3 more
- 11 Oct 2018
TL;DR: BERT as mentioned in this paper pre-trains deep bidirectional representations from unlabeled text by jointly conditioning on both left and right context in all layers, which can be fine-tuned with just one additional output layer to create state-of-the-art models for a wide range of tasks.
24.6K
Enriching Word Vectors with Subword Information
TL;DR: This paper proposed a new approach based on skip-gram model, where each word is represented as a bag of character n-grams, words being represented as the sum of these representations, allowing to train models on large corpora quickly and allowing to compute word representations for words that did not appear in the training data.
Representational Similarity Analysis – Connecting the Branches of Systems Neuroscience
TL;DR: A new experimental and data-analytical framework called representational similarity analysis (RSA) is proposed, in which multi-channel measures of neural activity are quantitatively related to each other and to computational theory and behavior by comparing RDMs.
3.6K
•Book
Basic Color Terms: Their Universality and Evolution
Paul Kay
- 01 Jan 1969
TL;DR: In this paper, the data, hypothesis, and general findings have been presented, including the evolution of basic color terms, and the data and hypothesis of the color term evolution, and some speculations.
2.9K