English
Related papers

Related papers: Training Transformers as a Universal Computer

200 papers

Recent works attribute the capability of in-context learning (ICL) in large pre-trained language models to implicitly simulating and fine-tuning an internal model (e.g., linear or 2-layer MLP) during inference. However, such constructions…

Computation and Language · Computer Science 2024-02-09 Abhishek Panigrahi , Sadhika Malladi , Mengzhou Xia , Sanjeev Arora

We consider computations of a Turing machine subjected to noise. In every step, the action (the new state and the new content of the observed cell, the direction of the head movement) can differ from that prescribed by the transition…

Computational Complexity · Computer Science 2021-12-07 Ilir Çapuni , Peter Gács

In reinforcement learning, wrappers are universally used to transform the information that passes between a model and an environment. Despite their ubiquity, no library exists with reasonable implementations of all popular preprocessing…

Machine Learning · Computer Science 2021-11-17 J. K. Terry , Benjamin Black , Ananth Hari

Transformers have become the dominant architecture for sequence modeling tasks such as natural language processing or audio processing, and they are now even considered for tasks that are not naturally sequential such as image…

Machine Learning · Computer Science 2024-03-05 Jorg Bornschein , Yazhe Li , Amal Rannen-Triki

Human language is known to exhibit a nested, hierarchical structure, allowing us to form complex sentences out of smaller pieces. However, many state-of-the-art neural networks models such as Transformers have no explicit hierarchical…

Computation and Language · Computer Science 2023-07-12 Nilay Patel , Jeffrey Flanigan

It is highly desirable for robots that work alongside humans to be able to understand instructions in natural language. Existing language conditioned imitation learning models directly predict the actuator commands from the image…

Robotics · Computer Science 2021-03-23 Sagar Gubbi Venkatesh , Raviteja Upadrashta , Bharadwaj Amrutur

Humans are excellent at understanding language and vision to accomplish a wide range of tasks. In contrast, creating general instruction-following embodied agents remains a difficult challenge. Prior work that uses pure language-only models…

Computer Vision and Pattern Recognition · Computer Science 2023-03-28 Hao Liu , Lisa Lee , Kimin Lee , Pieter Abbeel

A computational limit of combining partial evaluation and program inversion is investigated. Using a reversible Turing machine interpreter, we show that the first Futamura and inversion projections can produce not only functionally but also…

Programming Languages · Computer Science 2024-12-05 Robert Glück , Louis Marott Normann

The remarkable capability of Transformers to do reasoning and few-shot learning, without any fine-tuning, is widely conjectured to stem from their ability to implicitly simulate a multi-step algorithms -- such as gradient descent -- with…

Machine Learning · Computer Science 2024-10-14 Khashayar Gatmiry , Nikunj Saunshi , Sashank J. Reddi , Stefanie Jegelka , Sanjiv Kumar

This paper investigates the failure cases and out-of-distribution behavior of transformers trained on matrix inversion and eigenvalue decomposition. I show that incorrect model predictions still retain deep mathematical properties of the…

Machine Learning · Computer Science 2022-11-02 François Charton

Many tools used to process programs, like compilers, analyzers, or verifiers, perform transformations on their intermediate program representation, like abstract syntax trees. Implementing such program transformations is a non-trivial task,…

Programming Languages · Computer Science 2026-01-21 Michael Hanus , Steven Libby

Large-scale transformers achieve impressive results on program synthesis benchmarks, yet their true generalization capabilities remain obscured by data contamination and opaque training corpora. To rigorously assess whether models are truly…

Machine Learning · Computer Science 2026-05-01 Henrik Voigt , Michael Habeck , Joachim Giesen

In many real-world scenarios, data to train machine learning models becomes available over time. Unfortunately, these models struggle to continually learn new concepts without forgetting what has been learnt in the past. This phenomenon is…

Computation and Language · Computer Science 2023-01-16 Beyza Ermis , Giovanni Zappella , Martin Wistuba , Aditya Rawal , Cedric Archambeau

Transformers have achieved great success in natural language processing. Due to the powerful capability of self-attention mechanism in transformers, researchers develop the vision transformers for a variety of computer vision tasks, such as…

Computer Vision and Pattern Recognition · Computer Science 2022-07-08 Bo-Kai Ruan , Hong-Han Shuai , Wen-Huang Cheng

This paper describes a low-power processor tailored for fast Fourier transform computations where transport triggering template is exploited. The processor is software-programmable while retaining an energy-efficiency comparable to existing…

Hardware Architecture · Computer Science 2019-05-22 Jakub Žádník , Jarmo Takala

Supercompilation is a powerful program transformation technique with numerous interesting applications. Existing methods of supercompilation, however, are often very unpredictable with respect to the size of the resulting programs. We…

Programming Languages · Computer Science 2020-08-12 Dimitur Nikolaev Krustev

Initially developed for natural language processing (NLP), Transformers are now widely used for source code processing, due to the format similarity between source code and text. In contrast to natural language, source code is strictly…

Machine Learning · Computer Science 2021-06-25 Nadezhda Chirkova , Sergey Troshin

Trained transformer models have been found to implement interpretable procedures for tasks like arithmetic and associative recall, but little is understood about how the circuits that implement these procedures originate during training. To…

Machine Learning · Computer Science 2024-10-08 Ziqian Zhong , Jacob Andreas

We show that autoregressive decoding of a transformer-based language model can realize universal computation, without external intervention or modification of the model's weights. Establishing this result requires understanding how a…

Computation and Language · Computer Science 2024-10-07 Dale Schuurmans , Hanjun Dai , Francesco Zanini

This paper investigates the ability of transformer-based models to learn structural recursion from examples. Recursion is a universal concept in both natural and formal languages. Structural recursion is central to the programming language…

Computation and Language · Computer Science 2024-01-24 Dylan Zhang , Curt Tigges , Zory Zhang , Stella Biderman , Maxim Raginsky , Talia Ringer