---
title: README
emoji: 🏃
colorFrom: indigo
colorTo: gray
sdk: static
pinned: false
---

# Paris Noah's Ark Lab

## Projects

### Preprints

   - [Large Language Models as Markov Chains](https://huggingface.co./papers/2410.02724):  theoretical insights on their generalization and convergence properties.

### 2024

   - *(NeurIPS'24)* [MANO: Unsupervised Accuracy Estimation Under Distribution Shifts](https://huggingface.co./papers/2405.18979): when logits are enough to estimate generalization of a pre-trained model.
   - *(NeurIPS'24, **Spotlight**)* [Analysing Multi-Task Regression via Random Matrix Theory](https://arxiv.org/pdf/2406.10327): insights on a classical approach and its potentiality for time series forecasting.
   - *(ICML'24, **Oral**)* [SAMformer: Unlocking the Potential of Transformers in Time Series Forecasting](https://huggingface.co./papers/2402.10198): sharpness-aware minimization and channel-wise attention is all you need.
   - *(AISTATS'24)* [Leveraging Ensemble Diversity for Robust Self-Training](https://huggingface.co./papers/2310.14814): confidence estimation method for efficient pseudo-labeling under sample selection bias.
   - *(JMLR, 2024)* [Multi-class Probabilistic Bounds for Majority Vote Classifiers with Partially Labeled Data](https://www.jmlr.org/papers/volume25/23-0121/23-0121.pdf) generalization with unlabeled or pseudo-labeled data.
   - *(ICML '24)* [Position: A Call for Embodied AI](https://arxiv.org/abs/2402.03824): position paper on the need for embodied AI research
   - *(RLC '24)* [A Study of the Weighted Multi-step Loss Impact on the Predictive Error and the Return in MBRL](https://openreview.net/pdf?id=K4VjW7evSV): multi-step loss in MBRL does not work as well as expected