Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 7 of 7 for “"Length-Generalization"”.
-
Beyond Memorization: Evaluating Length-Generalization in Transformer-based Language Models
… However, transformers have trouble with length generalization (i.e., extrapolating on different lengths than seen during training). Recently, (Zhou et al., 2024a) proposed the RASP-Generalization Conjecture to predict what tasks are length-generalizable using a few carefully handcrafted …
-
Determination of output composition in reaction-advection-diffusion systems and improving language model performance with Re-Tuning
… improves performance on 3 common tasks in the length generalization literature.
-
Exploring Patterns of Algorithmic Generalization in Deep Learning
… deep learning systems ever exhibit algorithmic generalization? We investigate how deep learning systems can demonstrate algorithmic generalization, enabling them to handle problem instances of arbitrary size and complexity. We explore three patterns to achieve this goal: * Depth-wise recurrence: …
-
Stimulus control of human punished behavior: Testing for generalization.
… & Slivka, 1964) have investigated the stimulus generalization of punishment effects, or how they vary systematically with physical properties of stimuli. In a recent study with humans (O'Donnell & Crosbie, in press), each response in the presence of one horizontal line was punished, and …
-
On the Neural Architecture of Language Models
… memory, with emphasis on out-of-distribution length generalization. Empirically, PR-LSTM generalizes to longer sequences on more tasks than standard RNN, LSTM, and Transformer baselines. These results show that recurrent computation can be reorganized hierarchically to improve parallel …
-
Unifying Neural and Symbolic Computation for Compositional Generalization: Representation and Processing
… often fall short of the robust compositional generalization that characterizes human cognition, particularly in tasks demanding symbolic manipulation and algorithmic reasoning. This dissertation investigates the mechanisms underlying compositional generalization in neural networks and proposes …
-
Machine Learning-Based Data and Model Driven Bayesian Uncertanity Quantification of Inverse Problems for Suspended Non-structural System
… complexity measure, Minimum Description length (MDL), through Bayesian inference. It is validated in the SNS and yields optimal generalizability and computational scalability.