Trapped by simplicity: When Transformers fail to learn from noisy features

• URL: http://arxiv.org/abs/2602.08695v1
• Date: Mon, 09 Feb 2026 14:14:39 GMT
• Title: Trapped by simplicity: When Transformers fail to learn from noisy features
• Authors: Evan Peters, Ando Deng, Matheus H. Zambianco, Devin Blankespoor, Achim Kempf,
• Abstract summary: We study noise-robust learning: are transformers trained on data with noisy features able to find a target function that correctly predicts labels for noiseless features?<n>We find that transformers typically fail at noise-robust learning of random $k$-juntas, especially when the sensitivity of the optimal solution is smaller than that of the target function.
• Score: 0.0
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Noise is ubiquitous in data used to train large language models, but it is not well understood whether these models are able to correctly generalize to inputs generated without noise. Here, we study noise-robust learning: are transformers trained on data with noisy features able to find a target function that correctly predicts labels for noiseless features? We show that transformers succeed at noise-robust learning for a selection of $k$-sparse parity and majority functions, compared to LSTMs which fail at this task for even modest feature noise. However, we find that transformers typically fail at noise-robust learning of random $k$-juntas, especially when the boolean sensitivity of the optimal solution is smaller than that of the target function. We argue that this failure is due to a combination of two factors: transformers' bias toward simpler functions, combined with an observation that the optimal function for noise-robust learning typically has lower sensitivity than the target function for random boolean functions. We test this hypothesis by exploiting transformers' simplicity bias to trap them in an incorrect solution, but show that transformers can escape this trap by training with an additional loss term penalizing high-sensitivity solutions. Overall, we find that transformers are particularly ineffective for learning boolean functions in the presence of feature noise.

Related papers

• Transformers for Learning on Noisy and Task-Level Manifolds: Approximation and Generalization Insights [47.62295798627317]
  This work establishes a theoretical foundation by analyzing the performance of transformers for regression tasks involving noisy input data on a manifold.<n>We prove approximation and generalization errors which crucially depend on the intrinsic dimension of the manifold.<n>Our results demonstrate that transformers can leverage low-complexity structures in learning task even when the input data are perturbed by high-dimensional noise.
  arXiv  Detail & Related papers  (2025-05-06T05:41:46Z)
• One-Layer Transformer Provably Learns One-Nearest Neighbor In Context [48.4979348643494]
  We study the capability of one-layer transformers learning the one-nearest neighbor rule. A single softmax attention layer can successfully learn to behave like a one-nearest neighbor.
  arXiv  Detail & Related papers  (2024-11-16T16:12:42Z)
• Differential Transformer [99.5117269150629]
  Transformer tends to overallocate attention to irrelevant context.<n>We introduce Diff Transformer, which amplifies attention to relevant context while canceling noise.<n>It offers notable advantages in practical applications, such as long-context modeling, key information retrieval, hallucination mitigation, in-context learning, and reduction of activation outliers.
  arXiv  Detail & Related papers  (2024-10-07T17:57:38Z)
• Algorithmic Capabilities of Random Transformers [49.73113518329544]
  We investigate what functions can be learned by randomly transformers in which only the embedding layers are optimized. We find that these random transformers can perform a wide range of meaningful algorithmic tasks. Our results indicate that some algorithmic capabilities are present in transformers even before these models are trained.
  arXiv  Detail & Related papers  (2024-10-06T06:04:23Z)
• Learning on Transformers is Provable Low-Rank and Sparse: A One-layer Analysis [63.66763657191476]
  We show that efficient numerical training and inference algorithms as low-rank computation have impressive performance for learning Transformer-based adaption. We analyze how magnitude-based models affect generalization while improving adaption. We conclude that proper magnitude-based has a slight on the testing performance.
  arXiv  Detail & Related papers  (2024-06-24T23:00:58Z)
• Simplicity Bias in Transformers and their Ability to Learn Sparse Boolean Functions [29.461559919821802]
  Recent works have found that Transformers struggle to model several formal languages when compared to recurrent models. This raises the question of why Transformers perform well in practice and whether they have any properties that enable them to generalize better than recurrent models.
  arXiv  Detail & Related papers  (2022-11-22T15:10:48Z)
• Finetuning Pretrained Transformers into RNNs [81.72974646901136]
  Transformers have outperformed recurrent neural networks (RNNs) in natural language generation. A linear-complexity recurrent variant has proven well suited for autoregressive generation. This work aims to convert a pretrained transformer into its efficient recurrent counterpart.
  arXiv  Detail & Related papers  (2021-03-24T10:50:43Z)
• Revisiting Robust Neural Machine Translation: A Transformer Case Study [30.70732321809362]
  We investigate how noise breaks Transformers and if there exist solutions to deal with such issues. We introduce a novel data-driven technique to incorporate noise during training. We propose two new extensions to the original Transformer, that modify the neural architecture as well as the training process to handle noise.
  arXiv  Detail & Related papers  (2020-12-31T16:55:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.