#concept-frequency

[ follow ]
#pretraining
fromHackernoon
1 year ago
Artificial intelligence

Analyzing the Impact of Pretraining Frequency on Zero-Shot Performance in Multimodal Models | HackerNoon

Pretraining concept frequency is predictive of zero-shot performance across various multimodal models.
fromHackernoon
1 year ago
Artificial intelligence

AI Training Data Has a Long-Tail Problem | HackerNoon

Pretraining datasets exhibit a long-tailed distribution of concept frequencies, impacting performance disparities.
fromHackernoon
1 year ago
Artificial intelligence

Analyzing the Impact of Pretraining Frequency on Zero-Shot Performance in Multimodal Models | HackerNoon

more#pretraining
fromHackernoon
1 year ago

Across Metrics and Prompts, Frequent Concepts Outperform in Zero-Shot Learning | HackerNoon

The strong log-linear trend between concept frequency and zero-shot performance consistently holds across different prompting strategies, indicating that more frequently encountered concepts in pretraining data yield better performance.
Artificial intelligence
#pretraining-data
fromHackernoon
1 year ago
Artificial intelligence

AI Models Trained on Synthetic Data Still Follow Concept Frequency Trends | HackerNoon

fromHackernoon
1 year ago
Artificial intelligence

'Let It Wag!' and the Limits of Machine Learning on Rare Concepts | HackerNoon

fromHackernoon
1 year ago
Artificial intelligence

AI Models Trained on Synthetic Data Still Follow Concept Frequency Trends | HackerNoon

fromHackernoon
1 year ago
Artificial intelligence

'Let It Wag!' and the Limits of Machine Learning on Rare Concepts | HackerNoon

more#pretraining-data
Artificial intelligence
fromHackernoon
1 year ago

What 300GB of AI Research Reveals About the True Limits of "Zero-Shot" Intelligence | HackerNoon

Pretraining datasets impact the zero-shot performance of multimodal models through predictable frequency of concepts.
[ Load more ]