English
Related papers

Related papers: An empirical study on the limitation of Transforme…

200 papers

Effective trajectory generation is essential for reliable on-board spacecraft autonomy. Among other approaches, learning-based warm-starting represents an appealing paradigm for solving the trajectory generation problem, effectively…

Real-world processes often generate data that are a mix of categorical and numeric values that are recorded at irregular and informative intervals. Discrete token-based approaches are limited in numeric representation capacity while methods…

Machine Learning · Computer Science 2025-06-02 Andrew J. Loza , Jun Yup Kim , Shangzheng Song , Yihang Liu , Joseph J. Y. Sung , R Andrew Taylor , Dennis L. Shung

In-context learning enables transformer models to generalize to new tasks based solely on input prompts, without any need for weight updates. However, existing training paradigms typically rely on large, unstructured datasets that are…

Neural Processes (NPs) are a popular class of approaches for meta-learning. Similar to Gaussian Processes (GPs), NPs define distributions over functions and can estimate uncertainty in their predictions. However, unlike GPs, NPs and their…

Machine Learning · Computer Science 2023-02-09 Tung Nguyen , Aditya Grover

Empirical studies have identified a range of learnability biases and limitations of transformers, such as a persistent difficulty in learning to compute simple formal languages such as PARITY, and a bias towards low-degree functions.…

Machine Learning · Computer Science 2024-05-28 Michael Hahn , Mark Rofin

Programs with constraints are hard to debug. In this paper, we describe a general architecture to help develop new debugging tools for constraint programming. The possible tools are fed by a single general-purpose tracer. A tracer-driver is…

Software Engineering · Computer Science 2007-05-23 Ludovic Langevine , Mireille Ducasse

We show theoretically and empirically that the linear Transformer, when applied to graph data, can implement algorithms that solve canonical problems such as electric flow and eigenvector decomposition. The Transformer has access to…

Machine Learning · Computer Science 2025-03-04 Xiang Cheng , Lawrence Carin , Suvrit Sra

Large Language Models demonstrate remarkable mathematical capabilities but at the same time struggle with abstract reasoning and planning. In this study, we explore whether Transformers can learn to abstract and generalize the rules…

Neural and Evolutionary Computing · Computer Science 2024-12-03 Mikhail Burtsev

Transformer-based architectures have achieved remarkable success in natural language processing and computer vision. However, their performance in multivariate long-term forecasting often falls short compared to simpler linear baselines.…

Machine Learning · Computer Science 2025-07-09 Dizhen Liang

As state-of-the-art neural networks are deployed on reasoning and algorithmic tasks, exactness guarantees become increasingly important. However, high average-case accuracy can still mask inconsistent behaviors. This motivates exact…

Machine Learning · Computer Science 2026-05-25 Artur Back de Luca , Kimon Fountoulakis

Transformers have theoretical limitations in modeling certain sequence-to-sequence tasks, yet it remains largely unclear if these limitations play a role in large-scale pretrained LLMs, or whether LLMs might effectively overcome these…

Machine Learning · Computer Science 2025-10-24 Mayank Jobanputra , Yana Veitsman , Yash Sarrof , Aleksandra Bakalova , Vera Demberg , Ellie Pavlick , Michael Hahn

Sequence feature embedding is a challenging task due to the unstructuredness of sequence, i.e., arbitrary strings of arbitrary length. Existing methods are efficient in extracting short-term dependencies but typically suffer from…

Machine Learning · Statistics 2021-10-06 Chitta Ranjan , Samaneh Ebrahimi , Kamran Paynabar

Transformers have supplanted recurrent models in a large number of NLP tasks. However, the differences in their abilities to model different syntactic properties remain largely unknown. Past works suggest that LSTMs generalize very well on…

Computation and Language · Computer Science 2020-10-09 Satwik Bhattamishra , Kabir Ahuja , Navin Goyal

Recent work has shown that Transformers trained from scratch can successfully solve various arithmetic and algorithmic tasks, such as adding numbers and computing parity. While these Transformers generalize well on unseen inputs of the same…

Machine Learning · Computer Science 2025-05-13 Ying Fan , Yilun Du , Kannan Ramchandran , Kangwook Lee

One of the roadblocks for training generalist robotic models today is heterogeneity. Previous robot learning methods often collect data to train with one specific embodiment for one task, which is expensive and prone to overfitting. This…

Robotics · Computer Science 2024-10-01 Lirui Wang , Xinlei Chen , Jialiang Zhao , Kaiming He

In the burgeoning field of artificial intelligence (AI), understanding the capabilities and limitations of programming-oriented models is crucial. This paper presents a novel evaluation of the programming proficiency of Generative…

Artificial Intelligence · Computer Science 2023-06-06 Zizhuo Zhang , Lian Wen , Shaoyang Zhang , David Chen , Yanfei Jiang

We analyze the capabilities of Transformer language models in learning compositional discrete tasks. To this end, we evaluate training LLaMA models and prompting GPT-4 and Gemini on four tasks demanding to learn a composition of several…

The forecasting of entity trajectories at future points in time is a critical capability gap in applications across both Commercial and Defense sectors. Transformers, and specifically Generative Pre-trained Transformer (GPT) networks have…

Machine Learning · Computer Science 2024-02-02 Nicholas Stroh

Deep learning models such as the Transformer are often constructed by heuristics and experience. To provide a complementary foundation, in this work we study the following problem: Is it possible to find an energy function underlying the…

Machine Learning · Computer Science 2023-02-28 Yongyi Yang , Zengfeng Huang , David Wipf