Highlights of Ant Group’s 20 Accepted Papers at NeurIPS 2023
The article summarizes Ant Group's twenty accepted NeurIPS 2023 papers, covering advances in generative AI, time‑series forecasting, 3D image synthesis, and other machine‑learning topics, and provides brief overviews of three highlighted works along with links to the remaining studies.
On December 10, the NeurIPS 2023 conference in New Orleans received 12,343 submissions with a 26.1% acceptance rate, and Ant Group had 20 papers accepted across various AI and ML topics.
The highlighted papers include:
01. Prompt‑augmented Temporal Point Process for Streaming Event Sequence (PromptTPP) – introduces a prompt pool for continual learning in temporal point processes, addressing distribution shift and catastrophic forgetting in streaming event data.
02. BasisFormer: Attention‑based Time Series Forecasting with Learnable and Interpretable Basis – proposes a basis‑based forecasting model that uses cross‑attention to learn adaptive bases, achieving significant performance gains on multiple datasets.
03. Benchmarking and Analyzing 3D‑aware Image Synthesis with a Modularized Codebase – presents the Carver modular codebase that integrates NeRF with GANs, enabling systematic comparison of 3D image synthesis methods.
The article also lists the remaining 17 papers, covering topics such as optimizer design, diffusion models, causal inference, graph data augmentation, medical LLM evaluation, efficient pre‑training, and various generative‑adversarial techniques, each with authors and OpenReview links.
AntTech
Technology is the core driver of Ant's future creation.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.