Deep learning for source code modeling and generation: Models, applications, and challenges

THM Le, H Chen, MA Babar - ACM Computing Surveys (CSUR), 2020 - dl.acm.org
Deep Learning (DL) techniques for Natural Language Processing have been evolving
remarkably fast. Recently, the DL advances in language modeling, machine translation, and …

Wizardmath: Empowering mathematical reasoning for large language models via reinforced evol-instruct

H Luo, Q Sun, C Xu, P Zhao, J Lou, C Tao… - arxiv preprint arxiv …, 2023 - arxiv.org
Large language models (LLMs), such as GPT-4, have shown remarkable performance in
natural language processing (NLP) tasks, including challenging mathematical reasoning …

Virtualhome: Simulating household activities via programs

X Puig, K Ra, M Boben, J Li, T Wang… - Proceedings of the …, 2018 - openaccess.thecvf.com
In this paper, we are interested in modeling complex activities that occur in a typical
household. We propose to use programs, ie, sequences of atomic actions and interactions …

Leveraging grammar and reinforcement learning for neural program synthesis

R Bunel, M Hausknecht, J Devlin, R Singh… - arxiv preprint arxiv …, 2018 - arxiv.org
Program synthesis is the task of automatically generating a program consistent with a
specification. Recent years have seen proposal of a number of neural approaches for …

Recent advances in leveraging human guidance for sequential decision-making tasks

R Zhang, F Torabi, G Warnell, P Stone - Autonomous Agents and Multi …, 2021 - Springer
A longstanding goal of artificial intelligence is to create artificial agents capable of learning
to perform tasks that require sequential decision making. Importantly, while it is the artificial …

Execution-guided neural program synthesis

X Chen, C Liu, D Song - International Conference on Learning …, 2018 - openreview.net
Neural program synthesis from input-output examples has attracted an increasing interest
from both the machine learning and the programming language community. Most existing …

Neural program meta-induction

J Devlin, RR Bunel, R Singh… - Advances in Neural …, 2017 - proceedings.neurips.cc
Most recently proposed methods for Neural Program induction work under the assumption of
having a large set of input/output (I/O) examples for learning any given input-output …

Recursion of thought: A divide-and-conquer approach to multi-context reasoning with language models

S Lee, G Kim - arxiv preprint arxiv:2306.06891, 2023 - arxiv.org
Generating intermediate steps, or Chain of Thought (CoT), is an effective way to significantly
improve language models'(LM) multi-step reasoning capability. However, the CoT lengths …

Complex program induction for querying knowledge bases in the absence of gold programs

A Saha, GA Ansari, A Laddha… - Transactions of the …, 2019 - direct.mit.edu
Recent years have seen increasingly complex question-answering on knowledge bases
(KBQA) involving logical, quantitative, and comparative reasoning over KB subgraphs …

On-the-fly operation batching in dynamic computation graphs

G Neubig, Y Goldberg, C Dyer - Advances in Neural …, 2017 - proceedings.neurips.cc
Dynamic neural networks toolkits such as PyTorch, DyNet, and Chainer offer more flexibility
for implementing models that cope with data of varying dimensions and structure, relative to …