REVIEW 6 cited by
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
One Step of Gradient Descent is Provably the Optimal In-Context Learner with One Layer of Linear Self-Attention
read the original abstract
Recent works have empirically analyzed in-context learning and shown that transformers trained on synthetic linear regression tasks can learn to implement ridge regression, which is the Bayes-optimal predictor, given sufficient capacity [Aky\"urek et al., 2023], while one-layer transformers with linear self-attention and no MLP layer will learn to implement one step of gradient descent (GD) on a least-squares linear regression objective [von Oswald et al., 2022]. However, the theory behind these observations remains poorly understood. We theoretically study transformers with a single layer of linear self-attention, trained on synthetic noisy linear regression data. First, we mathematically show that when the covariates are drawn from a standard Gaussian distribution, the one-layer transformer which minimizes the pre-training loss will implement a single step of GD on the least-squares linear regression objective. Then, we find that changing the distribution of the covariates and weight vector to a non-isotropic Gaussian distribution has a strong impact on the learned algorithm: the global minimizer of the pre-training loss now implements a single step of $\textit{pre-conditioned}$ GD. However, if only the distribution of the responses is changed, then this does not have a large effect on the learned algorithm: even when the response comes from a more general family of $\textit{nonlinear}$ functions, the global minimizer of the pre-training loss still implements a single step of GD on a least-squares linear regression objective.
Forward citations
Cited by 6 Pith papers
-
Transformers Provably Learn to Internalize Chain-of-Thought
L-layer transformers under Log-ICoT curriculum provably learn k-parity with poly(n) samples and log k stages, matching explicit CoT efficiency without inference overhead.
-
Understanding In-Context Learning for Nonlinear Regression with Transformers: Attention as Featurizer
Transformers can be built to act as nonlinear featurizers via attention, supporting in-context regression with proven generalization bounds on synthetic tasks.
-
Learning to Adapt: In-Context Learning Beyond Stationarity
Gated linear attention enables lower training and test errors in non-stationary in-context learning by adaptively modulating past inputs through a learnable recency bias under an autoregressive model of task evolution.
-
Provable Knowledge Acquisition and Extraction in One-Layer Transformers
In a stylized one-layer transformer, pre-training encodes factual knowledge via relation-specific feature directions and attention patterns; fine-tuning extracts it through a relation-covering mechanism that succeeds ...
-
One for All: A Non-Linear Transformer can Enable Cross-Domain Generalization for In-Context Reinforcement Learning
Non-linear transformers enable cross-domain generalization in in-context RL by representing value functions from different domains with shared weights inside a shared RKHS.
-
A Survey on In-context Learning
The paper surveys definitions, techniques, applications, and challenges in in-context learning for large language models.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.