Published Mar 6, 2019

82 - Visual Reasoning, with Yoav Artzi

Yoav Artzi delves into the intricacies of visual reasoning by examining the balance between synthetic and natural datasets, discussing biases in model development, and exploring future challenges in enhancing AI's understanding of data representation. This episode highlights the evolving connection between language and visual stimuli, emphasizing the impact of sophisticated models and robust datasets on advancing AI comprehension.

Episode Highlights

Topics covered

Episode Highlights

Related Episodes

38 - A Corpus of Natural Language for Visual Reasoning, with Alane Suhr
Answers 383 questions
125 - VQA for Real Users, with Danna Gurari
Answers 383 questions
116 - Grounded Language Understanding, with Yonatan Bisk
Answers 383 questions
122 - Statutory Reasoning in Tax Law, with Nils Holzenberger
Answers 383 questions
91 - (Executable) Semantic Parsing, with Jonathan Berant
Answers 383 questions
95 - Common sense reasoning, with Yejin Choi
Answers 383 questions
108 - Data-To-Text Generation, with Verena Rieser and Ondřej Dušek
Answers 383 questions
20 - A simple neural network module for relational reasoning
Answers 383 questions
59 - Weakly Supervised Semantic Parsing With Abstract Examples, with Omer Goldman
Answers 383 questions
12 - Supervised Learning of Universal Sentence Representations from Natural Language Inference Data
Answers 383 questions
"Imaginative AI" with Mohamed Elhoseiny
Answers 383 questions
119 - Social NLP, with Diyi Yang
Answers 383 questions
25 - Neural Semantic Parsing over Multiple Knowledge-bases
Answers 383 questions
81 - BlackboxNLP, with Afra Alishahi and Tal Linzen
Answers 383 questions
130 - Linking human cognitive patterns to NLP Models, with Lisa Beinborn
Answers 383 questions

Dexa/NLP Highlights

82 - Visual Reasoning, with Yoav Artzi

Topics covered

Popular Clips

Natural Language Datasets

GQA Dataset Insights

Complex Reasoning in NLP

Linguistic Biases in AI

Visual Reasoning Challenges

Stimulus for Improvement

Image Reasoning Challenges

Neural Network Insights

Visual Reasoning Models

Visual Reasoning Insights

Data Set Highlights

Open Questions in Visual Reasoning

Classical vs. Neural Models

Episode Highlights

Diverse Datasets

Challenges in Modeling

Open Questions

Defining Visual Reasoning

Related Episodes

38 - A Corpus of Natural Language for Visual Reasoning, with Alane Suhr

125 - VQA for Real Users, with Danna Gurari

116 - Grounded Language Understanding, with Yonatan Bisk

122 - Statutory Reasoning in Tax Law, with Nils Holzenberger

91 - (Executable) Semantic Parsing, with Jonathan Berant

95 - Common sense reasoning, with Yejin Choi

108 - Data-To-Text Generation, with Verena Rieser and Ondřej Dušek

20 - A simple neural network module for relational reasoning

59 - Weakly Supervised Semantic Parsing With Abstract Examples, with Omer Goldman

12 - Supervised Learning of Universal Sentence Representations from Natural Language Inference Data

"Imaginative AI" with Mohamed Elhoseiny

119 - Social NLP, with Diyi Yang

25 - Neural Semantic Parsing over Multiple Knowledge-bases

81 - BlackboxNLP, with Afra Alishahi and Tal Linzen

130 - Linking human cognitive patterns to NLP Models, with Lisa Beinborn

82 - Visual Reasoning, with Yoav Artzi

Topics covered

Popular Clips

Episode Highlights

Diverse Datasets

Challenges in ModelingYoav Artzi examines the biases in visual reasoning datasets and the complexities involved in creating models that can handle intricate reasoning tasks. He highlights the importance of robust datasets and sophisticated models for advancing AI understanding.

Challenges in Modeling

Open QuestionsThe discussion shifts to the unresolved challenges and future directions in visual reasoning. Yoav Artzi and Matt Gardner explore the complexities of data representation and the potential for innovative approaches to enhance AI understanding.

Open Questions

Defining Visual ReasoningYoav Artzi explores the emerging field of visual reasoning, which connects language with visual stimuli to enable complex reasoning tasks. He discusses the significance of grounding language in visual contexts and the challenges posed by current datasets and models.

Defining Visual Reasoning

Related Episodes