As an Amazon Associate I earn from qualifying purchases.

A quick guide to Amazon’s papers at NeurIPS 2023

[ad_1]

The Conference on Neural Information Processing Systems (NeurIPS) takes place this week, and the Amazon papers accepted there touch on a wide range of topics, from experimental design and human-robot interaction to recommender systems and real-time statistical estimation. Amid that diversity, a few topics come in for particular attention: optimization, privacy, tabular data, time series forecasting, vision-language models — and particularly reinforcement learning.

Code generation

Large language models of code fail at completing code with potential bugs
Tuan Dinh, Jinman Zhao, Samson Tan, Renato Negrinho, Leonard Lausen, Sheng Zha, George Karypis

Complex query answering

Complex query answering on eventuality knowledge graph with implicit logical constraints
Jiaxin Bai, Xin Liu, Weiqi Wang, Chen Luo, Yangqiu Song

An example of a complex eventuality query, with its computational and informational atomics. V is something that happens before a person complains and leaves a restaurant; according to the knowledge graph, V could be either “Service is bad” or “Food is bad”. If V? is the reason for V, then according to the graph, V? could be either “Staff is new”, “PersonY adds ketchup”, “PersonY adds soy sauce”, or “PersonY adds vinegar”. However, from the query, we know that “PersonY adds vinegar” does not happen, and “PersonY adds soy sauce” happens after “food is bad”, so it can’t be the reason for “food is bad”. From “Complex query answering on eventuality knowledge graph with implicit logical constraints”.

Experimental design

Experimental designs for heteroskedastic variance
Justin Weltz, Tanner Fiez, Eric Laber, Alexander Volfovsky, Blake Mason, Houssam Nassif, Lalit Jain

Federated learning

Federated multi-objective learning
Haibo Yang, Zhuqing Liu, Jia Liu, Chaosheng Dong, Michinari Momma

Human-robot interaction

Alexa Arena: A user-centric interactive platform for embodied AI
Qiaozi (QZ) Gao, Govind Thattai, Suhaila Shakiah, Xiaofeng Gao, Shreyas Pansare, Vasu Sharma, Gaurav Sukhatme, Hangjie Shi, Bofei Yang, Desheng Zhang, Lucy Hu, Karthika Arumugam, Shui Hu, Matthew Wen, Dinakar Guthy, Cadence Chung, Rohan Khanna, Osman Ipek, Leslie Ball, Kate Bland, Heather Rocker, Michael Johnston, Reza Ghanadan, Dilek Hakkani-Tür, Prem Natarajan

Optimization

Bounce: Reliable high-dimensional Bayesian optimization for combinatorial and mixed spaces
Leonard Papenmeier, Luigi Nardi, Matthias Poloczek

Debiasing conditional stochastic optimization
Lie He, Shiva Kasiviswanathan

Distributionally robust Bayesian optimization with ϕ-divergences
Hisham Husain, Vu Nguyen, Anton van den Hengel

Ordinal classification

Conformal prediction sets for ordinal classification
Prasenjit Dey, Srujana Merugu, Sivaramakrishnan (Siva) Kaveri

Privacy

Creating a public repository for joining private data
James Cook, Milind Shyani, Nina Mishra

A stylized illustration of the repository problem. The sender S uploads a private count sketch capturing which people do and do not have cancer. The receiver R uses the sketch to decorate her data (people’s work locations) with a noisy version of S’s cancer column. Two noisy columns are generated: one for cancer (+1) and one for not (−1). R can then build a machine learning model to predict whether employees who work near a toxic waste site are more likely to develop cancer. From “Creating a public repository for joining private data”.

Scalable membership inference attacks via quantile regression
Martin Bertran Lopez, Shuai Tang, Michael Kearns, Jamie Morgenstern, Aaron Roth, Zhiwei Steven Wu

Real-time statistical estimation

Online robust non-stationary estimation
Abishek Sankararaman, Balakrishnan (Murali) Narayanaswamy

Recommender systems

Enhancing user intent capture in session-based recommendation with attribute patterns
Xin Liu, Zheng Li, Yifan Gao, Jingfeng Yang, Tianyu Cao, Zhengyang Wang, Bing Yin, Yangqiu Song

Reinforcement learning

Budgeting counterfactual for offline RL
Yao Liu, Pratik Chaudhari, Rasool Fakoor

Finite-time logarithmic Bayes regret upper bounds
Alexia Atsidakou, Branislav Kveton, Sumeet Katariya, Constantine Caramanis, Sujay Sanghavi

Resetting the optimizer in deep RL: An empirical study
Kavosh Asadi, Rasool Fakoor, Shoham Sabach

TD convergence: An optimization perspective
Kavosh Asadi, Shoham Sabach, Yao Liu, Omer Gottesman, Rasool Fakoor

Responsible AI

Improving fairness for spoken language understanding in atypical speech with text-to-speech
Helin Wang, Venkatesh Ravichandran, Milind Rao, Becky Lammers, Myra Sydnor, Nicholas Maragakis, Ankur A. Butala, Jayne Zhang, Victoria Chovaz, Laureano Moro-Velazquez

Tabular data

An inductive bias for tabular deep learning
Ege Beyazit, Jonathan Kozaczuk, Bo Li, Vanessa Wallace, Bilal Fadlallah

HYTREL: Hypergraph-enhanced tabular data representation learning
Pei Chen, Soumajyoti Sarkar, Leonard Lausen, Balasubramaniam Srinivasan, Sheng Zha, Ruihong Huang, George Karypis

An example of modeling a table as a hypergraph. Cells make up the nodes, and the cells in each row, each column, and the entire table form hyperedges. The table caption and the header names provide the names of the table and column hyperedges. The hypergraph keeps the four structural properties of tables — i.e., row/column permutations result in the same hypergraph. From “HYTREL: Hypergraph-enhanced tabular data representation learning”.

Time series forecasting

Predict, refine, synthesize: Self-guiding diffusion models for probabilistic time series forecasting
Marcel Kollovieh, Abdul Fatir Ansari, Michael Bohlke-Schneider, Jasper Zschiegner, Hao Wang, Yuyang (Bernie) Wang

PreDiff: Precipitation nowcasting with latent diffusion models
Zhihan Gao, Xingjian Shi, Boran Han, Hao Wang, Xiaoyong Jin, Danielle Maddix Robinson, Yi Zhu, Mu Li, Yuyang (Bernie) Wang

Vision-language models

Prompt pre-training with twenty-thousand classes for open-vocabulary visual recognition
Shuhuai Ren, Aston Zhang, Yi Zhu, Shuai Zhang, Shuai Zheng, Mu Li, Alex Smola, Xu Su

Your representations are in the network: Composable and parallel adaptation for large scale models
Yonatan Dukler, Alessandro Achille, Hao Yang, Ben Bowman, Varsha Vivek, Luca Zancato, Avinash Ravichandran, Charless Fowlkes, Ashwin Swaminathan, Stefano Soatto



[ad_2]

Source link

We will be happy to hear your thoughts

Leave a reply

Aqualib World- Space of Deals & offers
Logo