Search: openreview.net | Heykuki News

HK

Heykuki News

Top New Best Ask Show Jobs

Top New Best Ask Show Jobs

1.

Study identifies weaknesses in how AI systems are evaluated

7 months ago

416 points

2.

Launch HN: Humanloop (YC S20) – A platform to annotate, train and deploy NLP

6 years ago

157 points

3.

Launch HN: Roundtable (YC S23) – Using AI to Simulate Surveys

3 years ago

121 points

4.

Show HN: Open-Sourcing Google's Lattice Models in PyTorch

github.com/ControlAI

3 years ago

13 points

5.

Show HN: Roundtable – Estimating survey results in seconds

3 years ago

7 points

6.

Ask HN: Feasible to build open-source GATO Foundation Model?

3 years ago

3 points

7.

Ask HN: Some tips for evaluating intelligent agents?

2 years ago

2 points

8.

Show HN: ElasticMM – 4.2× Faster Multimodal LLM Serving (NeurIPS 2025 Oral)

6 months ago

1 points

9.

Show HN: Best Rejected Papers

deneutoy.github.io

a year ago

1 points

10.

DeltaProduct: Improving State-Tracking in Linear RNNs via Householder Products

a year ago

1 points

11.

Ask HN: Is there a mountain pick view concept in Reinforcement Learning?

4 years ago

1 points

12.

Measuring Compositional Generalization in ML Architectures

6 years ago

1 points

13.

30% drop in O1-preview accuracy when Putnam problems are slightly variated

a year ago

560 points

14.

Word2Vec received 'strong reject' four times at ICLR2013

3 years ago

369 points

15.

Deep Learning for Symbolic Mathematics

7 years ago

222 points

16.

Turing Complete Transformers: Two Transformers Are More Powerful Than One

2 years ago

190 points

17.

Vision Transformers Need Registers

2 years ago

171 points

18.

Transformers are inherently succinct

16 days ago

149 points

19.

Symbolic Regression is NP-hard

4 years ago

109 points

20.

Predictive Coding Approximates Backprop Along Arbitrary Computation Graphs

6 years ago

92 points

21.

Neural Architecture Search with Reinforcement Learning

9 years ago

88 points

22.

Approximating CNNs with Bag-of-Local-Features Models [pdf]

7 years ago

80 points

23.

A computational model of the Moth Olfactory Network learns to read MNIST [pdf]

8 years ago

71 points

24.

Text Embeddings Reveal (Almost) as Much as Text

3 years ago

66 points

25.

Large Language Models Are Human-Level Prompt Engineers

3 years ago

64 points

26.

ImageNet-trained CNNs are biased towards texture (2019)

6 years ago

60 points

27.

Learning to Represent Edits

7 years ago

39 points

28.

Learning the Arrow of Time [pdf]

7 years ago

36 points

29.

Learning to Design RNA

8 years ago

human_scientist

36 points

30.

Winner’s Curse? On pace, progress, and empirical rigor [pdf]

8 years ago

28 points