Google at ICLR 2022

Posted by Cat Armato and Callan Hajosy, Program Managers

The 10th International Conference on Learning Representations (ICLR 2022) kicks off this week, bringing together researchers, entrepreneurs, engineers and students alike to discuss and explore the rapidly advancing field of deep learning. Entirely virtual this year, ICLR 2022 offers conference and workshop tracks that present some of the latest research in deep learning and its applications to areas ranging from computer vision, speech recognition and text understanding to robotics, computational biology, and more.

As a Platinum Sponsor of ICLR 2022 and Champion DEI Action Fund contributor, Google will have a robust presence with nearly 100 accepted publications and extensive participation on organizing committees and in workshops. If you have registered for ICLR 2022, we hope you’ll watch our talks and learn about the work done at Google to address complex problems that affect billions of people. Here you can learn more about the research we will be presenting as well as our general involvement at ICLR 2022 (those with Google affiliations in bold).

Senior Area Chairs:
Includes: Been Kim, Dale Schuurmans, Sergey Levine

Area Chairs:
Includes: Adam White, Aditya Menon, Aleksandra Faust, Amin Karbasi, Amir Globerson, Andrew Dai, Balaji Lakshminarayanan, Behnam Neyshabur, Ben Poole, Bhuwan Dhingra, Bo Dai, Boqing Gong, Cristian Sminchisescu, David Ha, David Woodruff, Denny Zhou, Dipanjan Das, Dumitru Erhan, Dustin Tran, Emma Strubell, Eunsol Choi, George Dahl, George Tucker, Hanie Sedghi, Heinrich Jiang, Hossein Mobahi, Hugo Larochelle, Izhak Shafran, Jasper Snoek, Jean-Philippe Vert, Jeffrey Pennington, Justin Gilmer, Karol Hausman, Kevin Swersky, Krzysztof Choromanski, Mathieu Blondel, Matt Kusner, Michael Ryoo, Ming-Hsuan Yang, Minmin Chen, Mirella Lapata, Mohammad Ghavamzadeh, Mohammad Norouzi, Naman Agarwal, Nicholas Carlini, Olivier Bachem, Piyush Rai, Prateek Jain, Quentin Berthet, Richard Nock, Rose Yu, Sewoong Oh, Silvio Lattanzi, Slav Petrov, Srinadh Bhojanapalli, Tim Salimans, Ting Chen, Tong Zhang, Vikas Sindhwani, Weiran Wang, William Cohen, Xiaoming Liu

Workflow Chairs:
Includes: Yaguang Li

Diversity Equity & Inclusion Chairs:
Includes: Rosanne Liu

Invited Talks
Beyond Interpretability: Developing a Language to Shape Our Relationships with AI
Google Speaker: Been Kim

Do You See What I See? Large-Scale Learning from Multimodal Videos
Google Speaker: Cordelia Schmid

Publications
Hyperparameter Tuning with Renyi Differential Privacy – 2022 Outstanding Paper Award
Nicolas Papernot, Thomas Steinke

MIDI-DDSP: Detailed Control of Musical Performance via Hierarchical Modeling
Yusong Wu, Ethan Manilow, Yi Deng, Rigel Swavely, Kyle Kastner, Tim Cooijmans, Aaron Courville, Cheng-Zhi Anna Huang, Jesse Engel

The Information Geometry of Unsupervised Reinforcement Learning
Benjamin Eysenbach, Ruslan Salakhutdinov, Sergey Levine

Learning Strides in Convolutional Neural Networks – 2022 Outstanding Paper Award
Rachid Riad^*, Olivier Teboul, David Grangier, Neil Zeghidour

Poisoning and Backdooring Contrastive Learning
Nicholas Carlini, Andreas Terzis

Coordination Among Neural Modules Through a Shared Global Workspace
Anirudh Goyal, Aniket Didolkar, Alex Lamb, Kartikeya Badola, Nan Rosemary Ke, Nasim Rahaman, Jonathan Binas, Charles Blundell, Michael Mozer, Yoshua Bengio

Fine-Tuned Language Models Are Zero-Shot Learners (see the blog post)
Jason Wei, Maarten Bosma, Vincent Y. Zhao, Kelvin Guu, Adams Wei Yu, Brian Lester, Nan Du, Andrew M. Dai, Quoc V. Le

Large Language Models Can Be Strong Differentially Private Learners
Xuechen Li, Florian Tramèr, Percy Liang, Tatsunori Hashimoto

Progressive Distillation for Fast Sampling of Diffusion Models
Tim Salimans, Jonathan Ho

Exploring the Limits of Large Scale Pre-training
Samira Abnar, Mostafa Dehghani, Behnam Neyshabur, Hanie Sedghi

Scarf: Self-Supervised Contrastive Learning Using Random Feature Corruption
Dara Bahri, Heinrich Jiang, Yi Tay, Donald Metzler

Scalable Sampling for Nonsymmetric Determinantal Point Processes
Insu Han, Mike Gartrell, Jennifer Gillenwater, Elvis Dohmatob, Amin Karbasi

When Vision Transformers Outperform ResNets without Pre-training or Strong Data Augmentations
Xiangning Chen, Cho-Jui Hsieh, Boqing Gong

ViTGAN: Training GANs with Vision Transformers
Kwonjoon Lee, Huiwen Chang, Lu Jiang, Han Zhang, Zhuowen Tu, Ce Liu

Generalized Decision Transformer for Offline Hindsight Information Matching
Hiroki Furuta, Yutaka Matsuo, Shixiang Shane Gu

The MultiBERTs: BERT Reproductions for Robustness Analysis
Thibault Sellam, Steve Yadlowsky, Ian Tenney, Jason Wei, Naomi Saphra, Alexander D’Amour, Tal Linzen, Jasmijn Bastings, Iulia Turc, Jacob Eisenstein, Dipanjan Das, Ellie Pavlick

Scaling Laws for Neural Machine Translation
Behrooz Ghorbani, Orhan Firat, Markus Freitag, Ankur Bapna, Maxim Krikun, Xavier Garcia, Ciprian Chelba, Colin Cherry

Interpretable Unsupervised Diversity Denoising and Artefact Removal
Mangal Prakash, Mauricio Delbracio, Peyman Milanfar, Florian Jug

Understanding Latent Correlation-Based Multiview Learning and Self-Supervision: An Identifiability Perspective
Qi Lyu, Xiao Fu, Weiran Wang, Songtao Lu

Memorizing Transformers
Yuhuai Wu, Markus N. Rabe, DeLesley Hutchins, Christian Szegedy

Churn Reduction via Distillation
Heinrich Jiang, Harikrishna Narasimhan, Dara Bahri, Andrew Cotter, Afshin Rostamizadeh

DR3: Value-Based Deep Reinforcement Learning Requires Explicit Regularization
Aviral Kumar, Rishabh Agarwal, Tengyu Ma, Aaron Courville, George Tucker, Sergey Levine

Path Auxiliary Proposal for MCMC in Discrete Space
Haoran Sun, Hanjun Dai, Wei Xia, Arun Ramamurthy

On the Relation Between Statistical Learning and Perceptual Distances
Alexander Hepburn, Valero Laparra, Raul Santos-Rodriguez, Johannes Ballé, Jesús Malo

Possibility Before Utility: Learning And Using Hierarchical Affordances
Robby Costales, Shariq Iqbal, Fei Sha

MT3: Multi-Task Multitrack Music Transcription
Josh Gardner^*, Ian Simon, Ethan Manilow^*, Curtis Hawthorne, Jesse Engel

Bayesian Neural Network Priors Revisited
Vincent Fortuin, Adrià Garriga-Alonso, Sebastian W. Ober, Florian Wenzel, Gunnar Rätsch, Richard E. Turner, Mark van der Wilk, Laurence Aitchison

GradMax: Growing Neural Networks using Gradient Information
Utku Evci, Bart van Merrienboer, Thomas Unterthiner, Fabian Pedregosa, Max Vladymyrov

Scene Transformer: A Unified Architecture for Predicting Future Trajectories of Multiple Agents
Jiquan Ngiam, Benjamin Caine, Vijay Vasudevan, Zhengdong Zhang, Hao-Tien Lewis Chiang, Jeffrey Ling, Rebecca Roelofs, Alex Bewley, Chenxi Liu, Ashish Venugopal, David Weiss, Ben Sapp, Zhifeng Chen, Jonathon Shlens

The Role of Pretrained Representations for the OOD Generalization of RL Agents
Frederik Träuble, Andrea Dittadi, Manuel Wüthrich, Felix Widmaier, Peter Gehler, Ole Winther, Francesco Locatello, Olivier Bachem, Bernhard Schölkopf, Stefan Bauer

Autoregressive Diffusion Models
Emiel Hoogeboom, Alexey A. Gritsenko, Jasmijn Bastings, Ben Poole, Rianne van den Berg, Tim Salimans

The Role of Permutation Invariance in Linear Mode Connectivity of Neural Networks
Rahim Entezari, Hanie Seghi, Olga Saukh, Behnam Neyshabur

DISSECT: Disentangled Simultaneous Explanations via Concept Traversals
Asma Ghandeharioun, Been Kim, Chun-Liang Li, Brendan Jou, Brian Eoff, Rosalind W. Picard

Anisotropic Random Feature Regression in High Dimensions
Gabriel C. Mel, Jeffrey Pennington

Open-Vocabulary Object Detection via Vision and Language Knowledge Distillation
Xiuye Gu, Tsung-Yi Lin^*, Weicheng Kuo, Yin Cui

MCMC Should Mix: Learning Energy-Based Model with Flow-Based Backbone
Erik Nijkamp^*, Ruiqi Gao, Pavel Sountsov, Srinivas Vasudevan, Bo Pang, Song-Chun Zhu, Ying Nian Wu

Effect of Scale on Catastrophic Forgetting in Neural Networks
Vinay Ramasesh, Aitor Lewkowycz, Ethan Dyer

Incremental False Negative Detection for Contrastive Learning
Tsai-Shien Chen, Wei-Chih Hung, Hung-Yu Tseng, Shao-Yi Chien, Ming-Hsuan Yang

Towards Evaluating the Robustness of Neural Networks Learned by Transduction
Jiefeng Chen, Xi Wu, Yang Guo, Yingyu Liang, Somesh Jha

What Do We Mean by Generalization in Federated Learning?
Honglin Yuan^*, Warren Morningstar, Lin Ning, Karan Singhal

ViDT: An Efficient and Effective Fully Transformer-Based Object Detector
Hwanjun Song, Deqing Sun, Sanghyuk Chun, Varun Jampani, Dongyoon Han, Byeongho Heo, Wonjae Kim, Ming-Hsuan Yang

Measuring CLEVRness: Black-Box Testing of Visual Reasoning Models
Spyridon Mouselinos, Henryk Michalewski, Mateusz Malinowski

Wisdom of Committees: An Overlooked Approach To Faster and More Accurate Models (see the blog post)
Xiaofang Wang, Dan Kondratyuk, Eric Christiansen, Kris M. Kitani, Yair Alon (prev. Movshovitz-Attias), Elad Eban

Leveraging Unlabeled Data to Predict Out-of-Distribution Performance
Saurabh Garg^*, Sivaraman Balakrishnan, Zachary C. Lipton, Behnam Neyshabur, Hanie Sedghi

Data-Driven Offline Optimization for Architecting Hardware Accelerators (see the blog post)
Aviral Kumar, Amir Yazdanbakhsh, Milad Hashemi, Kevin Swersky, Sergey Levine

Diurnal or Nocturnal? Federated Learning of Multi-branch Networks from Periodically Shifting Distributions
Chen Zhu^*, Zheng Xu, Mingqing Chen, Jakub Konecny, Andrew Hard, Tom Goldstein

Policy Gradients Incorporating the Future
David Venuto, Elaine Lau, Doina Precup, Ofir Nachum

Discrete Representations Strengthen Vision Transformer Robustness
Chengzhi Mao^*, Lu Jiang, Mostafa Dehghani, Carl Vondrick, Rahul Sukthankar, Irfan Essa

SimVLM: Simple Visual Language Model Pretraining with Weak Supervision (see the blog post)
Zirui Wang, Jiahui Yu, Adams Wei Yu, Zihang Dai, Yulia Tsvetkov, Yuan Cao

Neural Stochastic Dual Dynamic Programming
Hanjun Dai, Yuan Xue, Zia Syed, Dale Schuurmans, Bo Dai

PolyLoss: A Polynomial Expansion Perspective of Classification Loss Functions
Zhaoqi Leng, Mingxing Tan, Chenxi Liu, Ekin Dogus Cubuk, Xiaojie Shi, Shuyang Cheng, Dragomir Anguelov

Information Prioritization Through Empowerment in Visual Model-Based RL
Homanga Bharadhwaj^*, Mohammad Babaeizadeh, Dumitru Erhan, Sergey Levine

Value Function Spaces: Skill-Centric State Abstractions for Long-Horizon Reasoning
Dhruv Shah, Peng Xu, Yao Lu, Ted Xiao, Alexander Toshev, Sergey Levine, Brian Ichter

Understanding and Leveraging Overparameterization in Recursive Value Estimation
Chenjun Xiao, Bo Dai, Jincheng Mei, Oscar Ramirez, Ramki Gummadi, Chris Harris, Dale Schuurmans

The Efficiency Misnomer
Mostafa Dehghani, Anurag Arnab, Lucas Beyer, Ashish Vaswani, Yi Tay

On the Role of Population Heterogeneity in Emergent Communication
Mathieu Rita, Florian Strub, Jean-Bastien Grill, Olivier Pietquin, Emmanuel Dupoux

No One Representation to Rule Them All: Overlapping Features of Training Methods
Raphael Gontijo-Lopes, Yann Dauphin, Ekin D. Cubuk

Data Poisoning Won’t Save You From Facial Recognition
Evani Radiya-Dixit, Sanghyun Hong, Nicholas Carlini, Florian Tramèr

AdaMatch: A Unified Approach to Semi-Supervised Learning and Domain Adaptation
David Berthelot, Rebecca Roelofs, Kihyuk Sohn, Nicholas Carlini, Alex Kurakin

Maximum Entropy RL (Provably) Solves Some Robust RL Problems
Benjamin Eysenbach, Sergey Levine

Auto-scaling Vision Transformers Without Training
Wuyang Chen, Wei Huang, Xianzhi Du, Xiaodan Song, Zhangyang Wang, Denny Zhou

Optimizing Few-Step Diffusion Samplers by Gradient Descent
Daniel Watson, William Chan, Jonathan Ho, Mohammad Norouzi

ExT5: Towards Extreme Multi-Task Scaling for Transfer Learning
Vamsi Aribandi, Yi Tay, Tal Schuster, Jinfeng Rao, Huaixiu Steven Zheng, Sanket Vaibhav Mehta, Honglei Zhuang, Vinh Q. Tran, Dara Bahri, Jianmo Ni, Jai Gupta, Kai Hui, Sebastian Ruder, Donald Metzler

Fortuitous Forgetting in Connectionist Networks
Hattie Zhou, Ankit Vani, Hugo Larochelle, Aaron Courville

Evading Adversarial Example Detection Defenses with Orthogonal Projected Gradient Descent
Oliver Bryniarski, Nabeel Hingun, Pedro Pachuca, Vincent Wang, Nicholas Carlini

Benchmarking the Spectrum of Agent Capabilities
Danijar Hafner

Charformer: Fast Character Transformers via Gradient-Based Subword Tokenization
Yi Tay, Vinh Q. Tran, Sebastian Ruder, Jai Gupta, Hyung Won Chung, Dara Bahri, Zhen Qin, Simon Baumgartner, Cong Yu, Donald Metzler

Mention Memory: Incorporating Textual Knowledge into Transformers Through Entity Mention Attention
Michiel de Jong, Yury Zemlyanskiy, Nicholas FitzGerald, Fei Sha, William Cohen

Eigencurve: Optimal Learning Rate Schedule for SGD on Quadratic Objectives with Skewed Hessian Spectrums
Rui Pan, Haishan Ye, Tong Zhang

Scale Efficiently: Insights from Pre-training and Fine-Tuning Transformers
Yi Tay, Mostafa Dehghani, Jinfeng Rao, William Fedus, Samira Abnar, Hyung Won Chung, Sharan Narang, Dani Yogatama, Ashish Vaswani, Donald Metzler

Omni-Scale CNNs: A Simple and Effective Kernel Size Configuration for Time Series Classification
Wensi Tang, Guodong Long, Lu Liu,Tianyi Zhou, Michael Blumenstein, Jing Jiang

Embedded-Model Flows: Combining the Inductive Biases of Model-Free Deep Learning and Explicit Probabilistic Modeling
Gianluigi Silvestri, Emily Fertig, Dave Moore, Luca Ambrogioni

Post Hoc Explanations May be Ineffective for Detecting Unknown Spurious Correlation
Julius Adebayo, Michael Muelly, Hal Abelson, Been Kim

Axiomatic Explanations for Visual Search, Retrieval, and Similarity Learning
Mark Hamilton, Scott Lundberg, Stephanie Fu, Lei Zhang, William T. Freeman

Pix2seq: A Language Modeling Framework for Object Detection (see the blog post)
Ting Chen, Saurabh Saxena, Lala Li, David J. Fleet, Geoffrey Hinton

Mirror Descent Policy Optimization
Manan Tomar, Lior Shani, Yonathan Efroni, Mohammad Ghavamzadeh

CodeTrek: Flexible Modeling of Code Using an Extensible Relational Representation
Pardis Pashakhanloo, Aaditya Naik, Yuepeng Wang, Hanjun Dai, Petros Maniatis, Mayur Naik

Conditional Object-Centric Learning From Video
Thomas Kipf, Gamaleldin F. Elsayed, Aravindh Mahendran, Austin Stone, Sara Sabour, Georg Heigold, Rico Jonschkowski, Alexey Dosovitskiy, Klaus Greff

A Loss Curvature Perspective on Training Instabilities of Deep Learning Models
Justin Gilmer, Behrooz Ghorbani, Ankush Garg, Sneha Kudugunta, Behnam Neyshabur, David Cardoze, George E. Dahl, Zack Nado, Orhan Firat

Autonomous Reinforcement Learning: Formalism and Benchmarking
Archit Sharma, Kelvin Xu, Nikhil Sardana, Abhishek Gupta, Karol Hausman, Sergey Levine, Chelsea Finn

TRAIL: Near-Optimal Imitation Learning with Suboptimal Data
Mengjiao Yang, Sergey Levine, Ofir Nachum

Minimax Optimization With Smooth Algorithmic Adversaries
Tanner Fiez, Lillian J. Ratliff, Chi Jin, Praneeth Netrapalli

Unsupervised Semantic Segmentation by Distilling Feature Correspondences
Mark Hamilton, Zhoutong Zhang, Bharath Hariharan, Noah Snavely, William T. Freeman

InfinityGAN: Towards Infinite-Pixel Image Synthesis
Chieh Hubert Lin, Hsin-Ying Lee, Yen-Chi Cheng, Sergey Tulyakov, Ming-Hsuan Yang

Shuffle Private Stochastic Convex Optimization
Albert Cheu, Matthew Joseph, Jieming Mao, Binghui Peng

Hybrid Random Features
Krzysztof Choromanski, Haoxian Chen, Han Lin, Yuanzhe Ma, Arijit Sehanobish, Deepali Jain, Michael S Ryoo, Jake Varley, Andy Zeng, Valerii Likhosherstov, Dmitry Kalashnikov, Vikas Sindhwani, Adrian Weller

Vector-Quantized Image Modeling With Improved VQGAN
Jiahui Yu, Xin Li, Jing Yu Koh, Han Zhang, Ruoming Pang, James Qin, Alexander Ku, Yuanzhong Xu, Jason Baldridge, Yonghui Wu

On the Benefits of Maximum Likelihood Estimation for Regression and Forecasting
Pranjal Awasthi, Abhimanyu Das, Rajat Sen, Ananda Theertha Suresh

Surrogate Gap Minimization Improves Sharpness-Aware Training
Juntang Zhuang^*, Boqing Gong, Liangzhe Yuan, Yin Cui, Hartwig Adam, Nicha C. Dvornek, Sekhar Tatikonda, James S. Duncan, Ting Liu

Online Target Q-learning With Reverse Experience Replay: Efficiently Finding the Optimal Policy for Linear MDPs
Naman Agarwal, Prateek Jain, Dheeraj Nagaraj, Praneeth Netrapalli, Syomantak Chaudhuri

CrossBeam: Learning to Search in Bottom-Up Program Synthesis
Kensen Shi, Hanjun Dai, Kevin Ellis, Charles Sutton

Workshops
Workshop on the Elements of Reasoning: Objects, Structure, and Causality (OSC)
Organizers include: Klaus Greff, Thomas Kipf

Workshop on Agent Learning in Open-Endedness
Organizers include: Krishna Srinivasan
Speakers include: Natasha Jaques, Danijar Hafner

Wiki-M3L: Wikipedia and Multi-modal & Multi-lingual Research
Organizers include: Klaus Greff, Thomas Kipf
Speakers include: Jason Baldridge, Tom Duerig

Setting Up ML Evaluation Standards to Accelerate Progress
Organizers include: Rishabh Agarwal
Speakers and Panelists include: Katherine Heller, Sara Hooker, Corinna Cortes

From Cells to Societies: Collective Learning Across Scales
Organizers include: Mark Sandler, Max Vladymyrov
Speakers include: Blaise Aguera y Arcas, Alexander Mordvintsev, Michael Mozer

Emergent Communication: New Frontiers
Speakers include: Natasha Jaques

Deep Learning for Code
Organizers include: Jonathan Herzig

GroundedML: Anchoring Machine Learning in Classical Algorithmic Theory
Speakers include: Gintare Karolina Dziugaite

Generalizable Policy Learning in the Physical World
Speakers and Panelists include: Mrinal Kalakrishnan

CoSubmitting Summer (CSS) Workshop
Organizers include: Rosanne Liu

^*Work done while at Google. ^↩

Source: Google AI Blog

Quick access to additional actions when composing a message in Google Chat on iOS

Quick launch summary

When using Google Chat on iOS, you can now easily take additional actions by hovering over the plus (“+”) icon next to the compose bar. You’ll see a variety of options such as:

Sharing a Google Meet link
Creating a meeting in Calendar
Accessing Google Drive
Text formatting options and more.

We hope this makes it easier to do your best work and collaborate when using Google Chat on your mobile device.

Getting started

Admins: There is no admin action required.
End users: Visit the Help Center to learn more about how to use Google Chat.

Rollout pace

Rapid Release and Scheduled Release domains: Extended rollout (potentially longer than 15 days for feature visibility) starting on April 25, 2022

Availability

Available to all Google Workspace customers and users with personal Google Accounts

Resources

Source: Google Workspace Updates

Easily manage storage related activity and policies through new storage management tools in the Admin console

What’s changing

We’re rolling out new storage management tools to give our customers additional visibility, control, and insights into storage usage across users, groups, and their entire organization.

In the Admin console, storage related activities can now be accessed and managed from a single source. Using these new storage management tools, admins can quickly and easily:

View a storage usage summary for their entire organization
View storage used by specific products like Drive or Gmail
View the top users of storage in their organization
View shared drives with the most storage used in your organization
Manage and delete shared drives based on storage use, including the ability to sort and delete individual or multiple shared drives
View storage limit warnings Access detailed reports on storage usage Apply storage limits for users

See below for more information.

Who’s impacted

Admins

Why it’s important

Admins can use these new tools to see how much storage is being used across their organization and view how close their organization is to reaching their storage limit.

In Storage Settings you can manage storage limits in Google Workspace across your organization. You’ll see the storage limit settings for your entire organization, which you can customize for specific organizational units and groups. Please visit the Help Center for more details on how storage is being used—and how you can manage it—across your organization.

This setting is turned OFF by default - when it’s turned on, you can create individual storage limits.

Easily modify storage privileges

At launch, only super admins will have access to the storage management tools. Over the coming months, access will expand to delegated, user, and reseller admins. When available, admins can control all previously implemented storage policies, allowing them to control defaults and create custom roles to manage storage policies for their organization or users in specific organizational units or groups.

Getting started

Super Admins: The new storage landing page can be accessed via:

The “Storage” option in the left-hand navigation menu.
A new “Storage” card on the Admin console homepage.
Or by navigating to Account > Settings > Storage.

Google Workspace customers: Visit the Help Center to learn more about storage in Google Workspace and managing shared drive users and their activity.
Google Workspace for Education customers: Visit the Help Center to learn more about Google Workspace for Education storage.
Important Note: At launch, super admins will have access to the storage management tools. We will share an update regarding access for delegated, user, and reseller admins on the Workspace Updates Blog once available.
End users: No action required.

Availability

Available for all Google Workspace super admins, as well as legacy G Suite Basic and Business super admins

Resources

Source: Google Workspace Updates

The urgent necessity of enacting a national privacy law

The following is adapted from remarks delivered by Kent Walker, President of Global Affairs, at Beyond the Basics: The Many Pillars of U.S. Privacy Law, an event hosted by R Street Institute at The National Press Club in Washington, DC. Google also published an accompanying white paperon Responsible Data Practices.

Information is all around us. Americans sometimes take it for granted, but from the moment we walk out our front doors, information powers everything we do.

After a two-years-and-counting pandemic, when people have taken to tech at an unprecedented pace, they’re more aware of both the possibilities and the privacy challenges.

They may have even heard about the shadowy world of data brokers who buy and sell information to actors they’ve never heard of, for purposes that they can’t see or control, in ways that may risk their privacy and security.

And they may have a greater appreciation for the need for consistency across the country — not a patchwork of 50 different state laws, but a law that organizations and people can rely on as they go about their daily lives

There is a range of views when it comes to technology and technology regulation. But when it comes to national privacy regulation, there is a clear consensus: Americans want it.

A Pew Research poll found that 75 percent of people support government regulation of consumer data.

And the absence of a comprehensive federal privacy law has left a vacuum that states are trying to fill by scrambling to pass their own, often inconsistent, laws — a trend that actually risks fragmenting consumer protections.

People are counting on all of us to address this issue — and fast. The good news is that after many years of discussion, today, there seems to be a growing consensus on this. We are starting to see interest from both parties, from many different constituencies. They are coming together on how to do this well.

President Biden in his State of the Union address highlighted the importance of privacy, and there are growing reports that Congress is making progress toward comprehensive privacy legislation. We’ve long supported that goal, and we welcome the forward movement.

When data is misused, when consumers find their trust is misplaced, it hurts not just the whole digital ecosystem, but the potential for future innovation.

And let me be clear: We at Google get it, and we’ve rethought and adapted our own approaches to product development to promote privacy and security.

For example, because digital services should keep your information for only as long as you find it helpful, we introduced auto-delete controls to let you easily delete your location history, web history, and YouTube history.

Try to do that with any other business that holds data about you.

We were the first platform to make it easy for people to download or transfer personal data when they want to switch to other services.

And today, we keep more people safe online than anyone else in the world — because if it’s not secure, it’s not private.

To set new standards for responsible data use, we’ve also done what we do best – built new technological solutions, investing in privacy-preserving technologies.

Privacy-preserving technologies don’t just promote privacy by design, they achieve privacy through innovation. They help us minimize the collection of identifying data. They reduce the risk of data being misused — without undermining the tremendous value that people get from information services.

As an example, at the start of COVID, we had an unprecedented partnership with Apple to develop Exposure Notifications, helping public health authorities supplement contact-tracing. Our North Star had to be designing a system with privacy protections baked in. So we worked with public health officials, privacy experts, regulators, used our most advanced technology to keep data safe, and established strict guidelines – all of which built public trust and adoption, saving thousands of lives.

Now we’ve got a complex business, and we haven’t always gotten everything right, but we’ve learned from those experiences, and we know what’s possible when private industry and regulators work together.

Of course it’s not enough for some organizations to operate responsibly — we need a law that establishes consistent rules and reins in bad actors.

So how do we do that? What’s the best path forward?

We're not focused on pie-in-the-sky proposals like creating an entirely new agency to regulate all the different uses of digital tools. We don’t want snappy soundbites; we want sound solutions.

The reality is that all companies are becoming digital companies, each with the potential to create new technologies and use information in new ways. We need consistent rules across the economy, and across the country.

Instead of chasing theoretical approaches, we want to support the practical, real-world privacy work already being done by Congress.

Current legislative privacy proposals like the ones put forward by Senators Cantwell and Wicker reflect important areas of agreement on the practical points that matter to people. And we hope they will work closely with Chairman Pallone and Ranking Member McMorris Rodgers to move legislation through the committees expeditiously.

We can build on the work that has already happened in this space, like proposals put forward by Senators Cortez Masto and Fischer and Representatives Stevens and Gonzalez to promote privacy-preserving technologies.

With the right leadership from the White House and leadership in Congress, we can get this done – this year.

So what are the sticking points? Issues like when and how consumers can file suit? The scope of FTC rulemaking? How federal and state laws will work together?

Those issues are debated in some form nearly every time Congress passes new business regulations, including the sectoral privacy laws Congress has already passed. So, none of this is new or unresolvable. With the right working group and some reasonable compromises, these points can be reconciled.

In fact, those conversations are already happening. Of course there has been no shortage of positions when it comes to privacy, ranging from ideas of notice and choice to proposals around new duties of care or loyalty.

One possible finesse would be a responsible data approach that works in practice, across a growing digital economy.

For example, we could start by giving consumers reasonable baseline assurances around transparency and control.

And we could build on that, by requiring responsible data practices — like privacy reviews and data minimization — that could be easy to implement and promote shared processes for protecting people’s data. Norms around good development processes could improve privacy practices for everyone.

But the time to act is now.

A U.S. privacy law would align us all on the privacy measures that people want and promote confidence in U.S. companies and our digital ecosystem.

It would increase trust in U.S. leadership, as we promote cross-border data flows and compatible, pro-privacy, pro-innovation rules around the world.

It would give everyone much-needed clarity and consistency so that organizations spend less time trying to navigate inconsistent rules and more time preventing harm and responsibly innovating – the kind of work that yields research breakthroughs and a stronger U.S. economy.

There’s no question that getting it done will take thoughtful compromises. Compromises by different groups in Congress. Compromises by advocates. And compromises by companies, including Google, who are used to doing business in certain ways. But that’s what we need to get this done.

Whatever final legislation comes out of the negotiations won’t be perfect, and it won’t address every concern. But we urge both businesses and advocates not to make the perfect the enemy of the good. Or of better, more consistent protections for all Americans.

In closing, I’ll say this: Google is an engineering company — and we look at problems from an engineering perspective. When we spot an issue with our services, we make fixing it a priority, and we often move engineers from other projects to help.

This is that all-hands-on-deck moment for privacy.

The vast majority of Americans want a federal privacy law. In fact, we’ve never seen such broad-based, bipartisan consensus about the need for that law.

It’s a moment for Congress to come together, on a bipartisan basis, and deliver for the American people.

Lawmakers and regulators face an important challenge, and an important opportunity. We pledge our support for that effort, and we hope that a broad cross-section of stakeholders will join together in support of their work.

Source: The Official Google Blog

A brief history of vaccination

Since at least the 1400s, people have looked for ways to protect themselves against infectious diseases. From the practice of “variolation” in the 15th century to today’s mRNA vaccines, immunization has a long history. Integral to that history has been the World Health Organization (WHO), whose global vaccine drives through the 20th and 21st centuries have played such a crucial role in reducing serious illness. For World Immunization Week, WHO has teamed up with Google Arts & Culture and scientific institutions from around the world to bring this history vividly to life with A Brief History of Vaccination.

From insufflation to vaccination

Looking back at the history of vaccination, with detailed stories drawn from medical archives, you’ll discover how we arrived at the jabs that have saved lives across the world. While you’ll encounter famous pioneers like Lady Mary Wortley Montagu, Edward Jenner and Louis Pasteur, you’ll also learn that vaccination has a much older history. In 15th-century China, for instance, there existed the practice of “insufflation” — blowing dried smallpox scabs into the nostril with a pipe to prevent natural smallpox, which was far more dangerous.

It was in the 20th century that earlier discoveries really started to bear fruit. Smallpox was eradicated globally and vaccines for polio, measles, influenza, hepatitis B, meningitis and many other diseases were developed. It was also the century that saw the inauguration of the WHO and its vital “Expanded Programme on Immunization,”which opened up a truly global front against vaccine-preventable diseases. A Brief History of Vaccination helps you to experience these great advances through photos, archive footage and historic scientific documents.

There are also those whose stories aren’t so well known, but nevertheless deserve to be told. You’ll learn about the enlightened Grand Duke of Tuscany who experimented with inoculation in the 18th century. Also featured here are the Mexican authorities whose efforts to defeat smallpox in the 19th century were ahead of their time.

Painting of a man inspecting a glass bottle while leaning on a table stacked with scientific instruments
Louis Pasteur in his laboratory, from the collection of Institut Pasteur
A photograph of two women talking outside a vaccination room holding a child
Vaccination campaign in Nigeria, from the collection of Fundación Princesa de Asturias
A painting of a man getting vaccinated with a syringe by a doctor, in a room full of people.
Vaccination, an introduction from the collection of National Council of Science Museum
A large painted poster with several small sections showing activities related to the Smallpox Eradication Programme
Ethiopian poster illustrating the activities of the Smallpox Eradication Programme, from the collection of World Health Organization

Unfinished history

Of course, the struggle against infectious disease is ongoing. During the COVID-19 pandemic, new stories emerged of ingenuity and resilience against the odds. You’ll learn of the heroism of Spanish and British health workers, and the man from Uttarakhand who became a one-man ambulance service in the remote mountain villages of northern India.

As authorities and communities around the world have strived to contain the pandemic, it has become ever more apparent that education is key to any successful vaccination program. With this in mind, educators can find a clear and accessible lesson plan that will provide learners with useful information about vaccination history.

Through A Brief History of Vaccination we learn, above all, that our fight against infectious diseases has united people across continents and cultures. As Louis Pasteur observed, “Science knows no country, because knowledge belongs to humanity, and is the torch which illuminates the world.”

Source: The Official Google Blog

Meet 11 startups working to combat climate change

We believe that technology and entrepreneurship can help avert the world’s climate crisis. Startup founders are using tools — from machine learning to mobile platforms to large scale data processing — to accelerate the change to a low-carbon economy. As part ofGoogle’s commitment to address climate change, we’ll continue to invest in the technologists and entrepreneurs who are working to build climate solutions.

So this Earth Day, we’re announcing the second Google for Startups Accelerator: Climate Change cohort. This ten-week program consists of intensive workshops and expert mentorship designed to help growth-stage, sustainability-focused startups learn technical, product and leadership best practices. Meet the 11 selected startups using technology to better our planet:

AmpUpin Cupertino, California: AmpUp is an electric vehicle (EV) software company and network provider that helps drivers, hosts, and fleets to charge stress-free.
Carbon Limitin Boca Raton, Florida: Carbon Limit transforms concrete into a CO2 sponge with green cement nanotechnology, turning roads and buildings into permanent CO2 solutions.
ChargeNet Stationsin Los Angeles, California: ChargeNet Stations aims to make charging accessible and convenient in all communities, preventing greenhouse gas emissions through use of PV + storage.
ChargerHelp!In Los Angeles, California: ChargerHelp! provides on-demand repair of electric vehicle charging stations, while also building out local workforces, removing barriers and creating economic mobility within all communities.
CO-Zin Boulder, Colorado: CO-Z accelerates electricity decarbonization and empowers renters, homeowners and businesses with advanced control, automated savings and power failure protection.
Community Energy Labsin Portland, Oregon: Community Energy Labs uses artificial intelligence to make smart energy management and decarbonization both accessible and affordable for community building owners.
Moment Energyin Vancouver, British Columbia: Moment Energy repurposes retired electric vehicle (EV) batteries to provide clean, affordable and reliable energy storage.
Mi Terroin City of Industry, California: Mi Terro is a synthetic biology and advanced material company that creates home compostable, plastic-alternative biomaterials made from plant-based agricultural waste.
Nithioin Washington, DC: Nithio is an AI-driven platform for clean energy investment that standardizes credit risk to catalyze capital to address climate change and achieve universal energy access.
Re Companyin New York City, New York: Re Company is a reusable packaging subscription service that supplies reuse systems with optimally designed containers and cycles them back into the supply chain at end of life.
Understoryin Pacific Grove, California: Understory rapidly monitors and quantifies discrete landscape changes to mitigate the effects of environmental change and deliver actionable information for land management, habitat conservation and climate risk assessment.

When the program kicks off this summer, startups will receive mentoring and technical support tailored to their business through a mix of one-to-one and one-to-many learning sessions, both remotely and in-person, from Google engineers and external experts. Stay tuned on Google for Startups social channels to see their experience unfold over the next three months.

Learn more about Google for Startups Accelerator here, and the latest on Google’s commitment to sustainability here.

Source: The Official Google Blog

Pix2Seq: A New Language Interface for Object Detection

Posted by Ting Chen and David Fleet, Research Scientists, Google Research, Brain Team

Object detection is a long-standing computer vision task that attempts to recognize and localize all objects of interest in an image. The complexity arises when trying to identify or localize all object instances while also avoiding duplication. Existing approaches, like Faster R-CNN and DETR, are carefully designed and highly customized in the choice of architecture and loss function. This specialization of existing systems has created two major barriers: (1) it adds complexity in tuning and training the different parts of the system (e.g., region proposal network, graph matching with GIOU loss, etc.), and (2), it can reduce the ability of a model to generalize, necessitating a redesign of the model for application to other tasks.

In “Pix2Seq: A Language Modeling Framework for Object Detection”, published at ICLR 2022, we present a simple and generic method that tackles object detection from a completely different perspective. Unlike existing approaches that are task-specific, we cast object detection as a language modeling task conditioned on the observed pixel inputs. We demonstrate that Pix2Seq achieves competitive results on the large-scale object detection COCO dataset compared to existing highly-specialized and well-optimized detection algorithms, and its performance can be further improved by pre-training the model on a larger object detection dataset. To encourage further research in this direction, we are also excited to release to the broader research community Pix2Seq’s code and pre-trained models along with an interactive demo.

Pix2Seq Overview
Our approach is based on the intuition that if a neural network knows where and what the objects in an image are, one could simply teach it how to read them out. By learning to “describe” objects, the model can learn to ground the descriptions on pixel observations, leading to useful object representations. Given an image, the Pix2Seq model outputs a sequence of object descriptions, where each object is described using five discrete tokens: the coordinates of the bounding box’s corners [y_min, x_min, y_max, x_max] and a class label.

Pix2Seq framework for object detection. The neural network perceives an image, and generates a sequence of tokens for each object, which correspond to bounding boxes and class labels.

With Pix2Seq, we propose a quantization and serialization scheme that converts bounding boxes and class labels into sequences of discrete tokens (similar to captions), and leverage an encoder-decoder architecture to perceive pixel inputs and generate the sequence of object descriptions. The training objective function is simply the maximum likelihood of tokens conditioned on pixel inputs and the preceding tokens.

Sequence Construction from Object Descriptions
In commonly used object detection datasets, images have variable numbers of objects, represented as sets of bounding boxes and class labels. In Pix2Seq, a single object, defined by a bounding box and class label, is represented as [y_min, x_min, y_max, x_max, class]. However, typical language models are designed to process discrete tokens (or integers) and are unable to comprehend continuous numbers. So, instead of representing image coordinates as continuous numbers, we normalize the coordinates between 0 and 1 and quantize them into one of a few hundred or thousand discrete bins. The coordinates are then converted into discrete tokens as are the object descriptions, similar to image captions, which in turn can then be interpreted by the language model. The quantization process is achieved by multiplying the normalized coordinate (e.g., y_min) by the number of bins minus one, and rounding it to the nearest integer (the detailed process can be found in our paper).

Quantization of the coordinates of the bounding boxes with different numbers of bins on a 480 × 640 image. With a small number of bins/tokens, such as 500 bins (∼1 pixel/bin), it achieves high precision even for small objects.

After quantization, the object annotations provided with each training image are ordered into a sequence of discrete tokens (shown below). Since the order of the objects does not matter for the detection task per se, we randomize the order of objects each time an image is shown during training. We also append an End of Sequence (EOS) token at the end as different images often have different numbers of objects, and hence sequence lengths.

The bounding boxes and class labels for objects detected in the image on the left are represented in the sequences shown on the right. A random object ordering strategy is used in our work but other approaches to ordering could also be used.

The Model Architecture, Objective Function, and Inference
We treat the sequences that we constructed from object descriptions as a “dialect” and address the problem via a powerful and general language model with an image encoder and autoregressive language encoder. Similar to language modeling, Pix2Seq is trained to predict tokens, given an image and preceding tokens, with a maximum likelihood loss. At inference time, we sample tokens from model likelihood. The sampled sequence ends when the EOS token is generated. Once the sequence is generated, we split it into chunks of 5 tokens for extracting and de-quantizing the object descriptions (i.e., obtaining the predicted bounding boxes and class labels). It is worth noting that both the architecture and loss function are task-agnostic in that they don’t assume prior knowledge about object detection (e.g., bounding boxes). We describe how we can incorporate task-specific prior knowledge with a sequence augmentation technique in our paper.

Results
Despite its simplicity, Pix2Seq achieves impressive empirical performance on benchmark datasets. Specifically, we compare our method with well established baselines, Faster R-CNN and DETR, on the widely used COCO dataset and demonstrate that it achieves competitive average precision (AP) results.

Pix2Seq achieves competitive AP results compared to existing systems that require specialization during model design, while being significantly simpler. The best performing Pix2Seq model achieved an AP score of 45.

Since our approach incorporates minimal inductive bias or prior knowledge of the object detection task into the model design, we further explore how pre-training the model using the large-scale object detection COCO dataset can impact its performance. Our results indicate that this training strategy (along with using bigger models) can further boost performance.

The average precision of the Pix2Seq model with pre-training followed by fine-tuning. The best performing Pix2Seq model without pre-training achieved an AP score of 45. When the model is pre-trained, we see an 11% improvement with an AP score of 50.

Pix2Seq can detect objects in densely populated and complex scenes, such as those shown below.

Example complex and densely populated scenes labeled by a trained Pix2Seq model. Try it out here.

Conclusion and Future Work
With Pix2Seq, we cast object detection as a language modeling task conditioned on pixel inputs for which the model architecture and loss function are generic, and have not been engineered specifically for the detection task. One can, therefore, readily extend this framework to different domains or applications, where the output of the system can be represented by a relatively concise sequence of discrete tokens (e.g., keypoint detection, image captioning, visual question answering), or incorporate it into a perceptual system supporting general intelligence, for which it provides a language interface to a wide range of vision and language tasks. We also hope that the release of our Pix2Seq’s code, pre-trained models and interactive demo will inspire further research in this direction.

Acknowledgements
This post reflects the combined work with our co-authors: Saurabh Saxena, Lala Li, Geoffrey Hinton. We would also like to thank Tom Small for the visualization of the Pix2Seq illustration figure.