Picture for Maxat Tezekbayev

Maxat Tezekbayev

Gradient Descent Fails to Learn High-frequency Functions and Modular Arithmetic

Add code
Oct 19, 2023
Viaarxiv icon

Intractability of Learning the Discrete Logarithm with Gradient-Based Methods

Add code
Oct 02, 2023
Viaarxiv icon

Long-Tail Theory under Gaussian Mixtures

Add code
Jul 24, 2023
Viaarxiv icon

Autoencoders for a manifold learning problem with a Jacobian rank constraint

Add code
Jun 25, 2023
Viaarxiv icon

Speeding Up Entmax

Add code
Nov 15, 2021
Figure 1 for Speeding Up Entmax
Figure 2 for Speeding Up Entmax
Figure 3 for Speeding Up Entmax
Figure 4 for Speeding Up Entmax
Viaarxiv icon

The Rediscovery Hypothesis: Language Models Need to Meet Linguistics

Add code
Mar 02, 2021
Figure 1 for The Rediscovery Hypothesis: Language Models Need to Meet Linguistics
Figure 2 for The Rediscovery Hypothesis: Language Models Need to Meet Linguistics
Figure 3 for The Rediscovery Hypothesis: Language Models Need to Meet Linguistics
Figure 4 for The Rediscovery Hypothesis: Language Models Need to Meet Linguistics
Viaarxiv icon

Semantics- and Syntax-related Subvectors in the Skip-gram Embeddings

Add code
Dec 23, 2019
Figure 1 for Semantics- and Syntax-related Subvectors in the Skip-gram Embeddings
Figure 2 for Semantics- and Syntax-related Subvectors in the Skip-gram Embeddings
Figure 3 for Semantics- and Syntax-related Subvectors in the Skip-gram Embeddings
Figure 4 for Semantics- and Syntax-related Subvectors in the Skip-gram Embeddings
Viaarxiv icon