Black-box machine learning models have gained widespread deployment in decision-making settings across many parts of society, from sentencing decisions to medical diagnostics to loan lending. However, many models were found to be biased against certain demographic groups. Initial work on Algorithmic fairness focused on formalizing statistical measures of fairness, that could be used to train new classifiers. While these models were an important first step towards addressing fairness concerns, there were immediate challenges with them. Causality has recently emerged as a powerful tool to address these shortcomings. Causality can be seen as a model-first approach: starting with the language of structural causal models or potential outcomes, the idea is to frame, then solve questions of algorithmic fairness in this language. Such causal definitions of fairness can have far-reaching impact, especially in high risk domains. Interpretability on the other hand can be viewed as a user-first approach: can the ways in which algorithms work be made more transparent, making it easier for them to align with our societal values on fairness? In this way, Interpretability can sometimes be more actionable than Causality work.

Given these initial successes, this workshop aims to more deeply investigate how open questions in algorithmic fairness can be addressed with Causality and Interpretability. Questions such as: What improvements can causal definitions provide compared to existing statistical definitions of fairness? How can causally grounded methods help develop more robust fairness algorithms in practice? What tools for interpretability are useful for detecting bias and building fair systems? What are good formalizations of interpretability when addressing fairness questions?

Website: www.afciworkshop.org

'Medical Imaging meets NeurIPS' is a satellite workshop established in 2017. The workshop aims to bring researchers together from the medical image computing and machine learning communities. The objective is to discuss the major challenges in the field and opportunities for joining forces. This year the workshop will feature online oral and poster sessions with an emphasis on audience interactions. In addition, there will be a series of high-profile invited speakers from industry, academia, engineering and medical sciences giving an overview of recent advances, challenges, latest technology and efforts for sharing clinical data.

Medical imaging is facing a major crisis with an ever increasing complexity and volume of data and immense economic pressure. The interpretation of medical images pushes human abilities to the limit with the risk that critical patterns of disease go undetected. Machine learning has emerged as a key technology for developing novel tools in computer aided diagnosis, therapy and intervention. Still, progress is slow compared to other fields of visual recognition which is mainly due to the domain complexity and constraints in clinical applications which require most robust, accurate, and reliable solutions. The workshop aims to raise the awareness of the unmet needs in machine learning for successful applications in medical imaging.

We propose to organize a workshop on machine learning and combinatorial algorithms. The combination of methods from machine learning and classical AI is an emerging trend. Many researchers have argued that “future AI” methods somehow need to incorporate discrete structures and symbolic/algorithmic reasoning. Additionally, learning-augmented optimization algorithms can impact the broad range of difficult but impactful optimization settings. Coupled learning and combinatorial algorithms have the ability to impact real-world settings such as hardware & software architectural design, self-driving cars, ridesharing, organ matching, supply chain management, theorem proving, and program synthesis among many others. We aim to present diverse perspectives on the integration of machine learning and combinatorial algorithms.

This workshop aims to bring together academic and industrial researchers in order to describe recent advances and build lasting communication channels for the discussion of future research directions pertaining the integration of machine learning and combinatorial algorithms. The workshop will connect researchers with various relevant backgrounds, such as those working on hybrid methods, have particular expertise in combinatorial algorithms, work on problems whose solution likely requires new approaches, as well as everyone interested in learning something about this emerging field of research. We aim to highlight open problems in bridging the gap between machine learning and combinatorial optimization in order to facilitate new research directions.
The workshop will foster the collaboration between the communities by curating a list of problems and challenges to promote the research in the field.

Our technical topics of interest include (but are not limited to):
- Hybrid architectures with combinatorial building blocks
- Attacking hard combinatorial problems with learning
- Neural architectures mimicking combinatorial algorithms

Further information about speakers, paper submissions and schedule are available at the workshop website: https://sites.google.com/view/lmca2020/home .

A few months ago, the world was shaken by the outbreak of the novel Coronavirus, exposing the lack of preparedness for such a case in many nations around the globe. As we watched the daily number of cases of the virus rise exponentially, and governments scramble to design appropriate policies, communities collectively asked “Could we have been better prepared for this?” Similar questions have been brought up by the climate emergency the world is now facing.
At a time of global reckoning, this year’s ML4D program will focus on building and improving resilience in developing regions through machine learning. Past iterations of the workshop have explored how machine learning can be used to tackle global development challenges, the potential benefits of such technologies, as well as the associated risks and shortcomings. This year we seek to ask our community to go beyond solely tackling existing problems by building machine learning tools with foresight, anticipating application challenges, and providing sustainable, resilient systems for long-term use.
This one-day workshop will bring together a diverse set of participants from across the globe. Attendees will learn about how machine learning tools can help enhance preparedness for disease outbreaks, address the climate crisis, and improve countries’ ability to respond to emergencies. It will also discuss how naive “tech solutionism” can threaten resilience by posing risks to human rights, enabling mass surveillance, and perpetuating inequalities. The workshop will include invited talks, contributed talks, a poster session of accepted papers, breakout sessions tailored to the workshop’s theme, and panel discussions.

Reinforcement learning (RL) algorithms learn through rewards and a process of trial-and-error. This approach is strongly inspired by the study of animal behaviour and has led to outstanding achievements. However, artificial agents still struggle with a number of difficulties, such as learning in changing environments and over longer timescales, states abstractions, generalizing and transferring knowledge. Biological agents, on the other hand, excel at these tasks. The first edition of our workshop last year brought together leading and emerging researchers from Neuroscience, Psychology and Machine Learning to share how neural and cognitive mechanisms can provide insights for RL research and how machine learning advances can further our understanding of brain and behaviour. This year, we want to build on the success of our previous workshop, by expanding on the challenges that emerged and extending to novel perspectives. The problem of state and action representation and abstraction emerged quite strongly last year, so this year’s program aims to add new perspectives like hierarchical reinforcement learning, structure learning and their biological underpinnings. Additionally, we will address learning over long timescales, such as lifelong learning or continual learning, by including views from synaptic plasticity and developmental neuroscience. We are hoping to inspire and further develop connections between biological and artificial reinforcement learning by bringing together experts from all sides and encourage discussions that could help foster novel solutions for both communities.

We’ve all been there. A creative spark leads to a beautiful idea. We love the idea, we nurture it, and name it. The idea is elegant: all who hear it fawn over it. The idea is justified: all of the literature we have read supports it. But, lo and behold: once we sit down to implement the idea, it doesn’t work. We check our code for software bugs. We rederive our derivations. We try again and still, it doesn’t work. We Can’t Believe It’s Not Better [1].

In this workshop, we will encourage probabilistic machine learning researchers who Can’t Believe It’s Not Better to share their beautiful idea, tell us why it should work, and hypothesize why it does not in practice. We also welcome work that highlights pathologies or unexpected behaviors in well-established practices. This workshop will stress the quality and thoroughness of the scientific procedure, promoting transparency, deeper understanding, and more principled science.

Focusing on the probabilistic machine learning community will facilitate this endeavor, not only by gathering experts that speak the same language, but also by exploiting the modularity of probabilistic framework. Probabilistic machine learning separates modeling assumptions, inference, and model checking into distinct phases [2]; this facilitates criticism when the final outcome does not meet prior expectations. We aim to create an open-minded and diverse space for researchers to share unexpected or negative results and help one another improve their ideas.

For full details see: [ protected link dropped ]

Modern engineering workflows are built on computational tools for specifying models and designs, for numerical analysis of system behavior, and for optimization, model-fitting and rational design. How can machine learning be used to empower the engineer and accelerate this workflow? We wish to bring together machine learning researchers and engineering academics to address the problem of developing ML tools which benefit engineering modeling, simulation and design, through reduction of required computational or human effort, through permitting new rich design spaces, through enabling production of superior designs, or through enabling new modes of interaction and new workflows.

Generative machine learning and machine creativity have continued to grow and attract a wider audience to machine learning. Generative models enable new types of media creation across images, music, and text - including recent advances such as StyleGAN2, Jukebox and GPT-3. This one-day workshop broadly explores issues in the applications of machine learning to creativity and design. We will look at algorithms for generation and creation of new media, engaging researchers building the next generation of generative models (GANs, RL, etc). We investigate the social and cultural impact of these new models, engaging researchers from HCI/UX communities and those using machine learning to develop new creative tools. In addition to covering the technical advances, we also address the ethical concerns ranging from the use of biased datasets to replicating artistic work. Finally, we’ll hear from some of the artists and musicians who are adopting machine learning including deep learning and reinforcement learning as part of their own artistic process. We aim to balance the technical issues and challenges of applying the latest generative models to creativity and design with philosophical and cultural issues that surround this area of research.

https://www.CooperativeAI.com/

Problems of cooperation—in which agents seek ways to jointly improve their welfare—are ubiquitous and important. They can be found at all scales ranging from our daily routines—such as highway driving, communication via shared language, division of labor, and work collaborations—to our global challenges—such as disarmament, climate change, global commerce, and pandemic preparedness. Arguably, the success of the human species is rooted in our ability to cooperate, in our social intelligence and skills. Since machines powered by artificial intelligence and machine learning are playing an ever greater role in our lives, it will be important to equip them with the skills necessary to cooperate and to foster cooperation.

We see an opportunity for the field of AI, and particularly machine learning, to explicitly focus effort on this class of problems which we term Cooperative AI. The goal of this research would be to study the many aspects of the problem of cooperation, and innovate in AI to contribute to solving these problems. Central questions include how to build machine agents with the capabilities needed for cooperation, and how advances in machine learning can help foster cooperation in populations of agents (of machines and/or humans), such as through improved mechanism design and mediation.

Research could be organized around key capabilities necessary for cooperation, including: understanding other agents, communicating with other agents, constructing cooperative commitments, and devising and negotiating suitable bargains and institutions. Since artificial agents will often act on behalf of particular humans and in ways that are consequential for humans, this research will need to consider how machines can adequately learn human preferences, and how best to integrate human norms and ethics into cooperative arrangements.

We are planning to bring together scholars from diverse backgrounds to discuss how AI research can contribute to the field of cooperation.


Call for Papers
We invite high-quality paper submissions on the following topics (broadly construed, this is not an exhaustive list):

-Multi-agent learning
-Agent cooperation
-Agent communication
-Resolving commitment problems
-Agent societies, organizations and institutions
-Trust and reputation
-Theory of mind and peer modelling
-Markets, mechanism design and and economics based cooperation
-Negotiation and bargaining agents
-Team formation problems

Accepted papers will be presented during joint virtual poster sessions and be made publicly available as non archival reports, allowing future submissions to archival conferences or journals.

Submissions should be up to eight pages excluding references, acknowledgements, and supplementary material, and should follow NeurIPS format. The review process will be double-blind.

Paper submissions: https://easychair.org/my/conference?conf=coopai2020#

Discovering new molecules and materials is a central pillar of human well-being, providing new medicines, securing the world’s food supply via agrochemicals, or delivering new battery or solar panel materials to mitigate climate change. However, the discovery of new molecules for an application can often take up to a decade, with costs spiraling. Machine learning can help to accelerate the discovery process. The goal of this workshop is to bring together researchers interested in improving applications of machine learning for chemical and physical problems and industry experts with practical experience in pharmaceutical and agricultural development. In a highly interactive format, we will outline the current frontiers and present emerging research directions. We aim to use this workshop as an opportunity to establish a common language between all communities, to actively discuss new research problems, and also to collect datasets by which novel machine learning models can be benchmarked. The program is a collection of invited talks, alongside contributed posters. A panel discussion will provide different perspectives and experiences of influential researchers from both fields and also engage open participant conversation. An expected outcome of this workshop is the interdisciplinary exchange of ideas and initiation of collaboration.

Following growing concerns with both harmful research impact and research conduct in computer science, including concerns with research published at NeurIPS, this year’s conference introduced two new mechanisms for ethical oversight: a requirement that authors include a “broader impact statement” in their paper submissions and additional evaluation criteria asking paper reviewers to identify any potential ethical issues with the submissions.

These efforts reflect a recognition that existing research norms have failed to address the impacts of AI research, and take place against the backdrop of a larger reckoning with the role of AI in perpetuating injustice. The changes have been met with both praise and criticism some within and outside the community see them as a crucial first step towards integrating ethical reflection and review into the research process, fostering necessary changes to protect populations at risk of harm. Others worry that AI researchers are not well placed to recognize and reason about the potential impacts of their work, as effective ethical deliberation may require different expertise and the involvement of other stakeholders.

This debate reveals that even as the AI research community is beginning to grapple with the legitimacy of certain research questions and critically reflect on its research practices, there remains many open questions about how to ensure effective ethical oversight. This workshop therefore aims to examine how concerns with harmful impacts should affect the way the research community develops its research agendas, conducts its research, evaluates its research contributions, and handles the publication and dissemination of its findings. This event complements other NeurIPS workshops this year devoted to normative issues in AI and builds on others from years past, but adopts a distinct focus on the ethics of research practice and the ethical obligations of researchers.

Is backpropagation the ultimate tool on the path to achieving synthetic intelligence as its success and widespread adoption would suggest?

Many have questioned the biological plausibility of backpropagation as a learning mechanism since its discovery. The weight transport and timing problems are the most disputable. The same properties of backpropagation training also have practical consequences. For instance, backpropagation training is a global and coupled procedure that limits the amount of possible parallelism and yields high latency.

These limitations have motivated us to discuss possible alternative directions. In this workshop, we want to promote such discussions by bringing together researchers from various but related disciplines, and to discuss possible solutions from engineering, machine learning and neuroscientific perspectives.

Public health and population health refer to the study of daily life factors and prevention efforts, and their effects on the health of populations. We expect that work featured in this workshop will differ from Machine Learning in Healthcare as it will focus on data and algorithms related to the non-medical conditions that shape our health including structural, lifestyle, policy, social, behavior and environmental factors. Indeed, much of the data that is traditionally used in machine learning and health problems are really about our interactions with the health care system, and this workshop aims to balance this with machine learning work using data on the non-medical conditions that shape our health. There are many machine learning opportunities specific to these data and how they are used to assess and understand health and disease, that differ from healthcare specific data and tasks (e.g. the data is often unstructured, must be captured across the life-course, in different environments, etc.) This is pertinent for both infectious diseases such as COVID-19 and non-communicable diseases such as diabetes, stroke, etc. Indeed, this workshop topic is especially timely given the COVID outbreak, protests regarding racism, and associated interest in exploring relevance of machine learning to questions around disease incidence, prevention and mitigation related to both of these and their synergy. These questions require the use of data from outside of healthcare, as well as considerations of how machine learning can augment work in epidemiology and biostatistics.

This workshop will focus on exploring the utility of interactive narratives to fill a role as the learning environments of choice for language-based tasks including but not limited to storytelling. A previous iteration of this workshop took place very successfully with over a hundred attendees, also at NeurIPS, in 2018 and since then the community of people working in this area has rapidly increased. This workshop aims to be a centralized place where all researchers involved across a breadth of fields can interact and learn from each other. Furthermore, it will act as a showcase to the wider NLP/RL/Game communities on interactive narrative's place as a learning environment. The program will feature a collection of invited talks in addition to contributed talks and posters from each of these sections of the interactive narrative community and the wider NLP and RL communities.

Over the last decade, deep networks have propelled machine learning to accomplish tasks previously considered far out of reach, human-level performance in image classification and game-playing. However, research has also shown that the deep networks are often brittle to distributional shifts in data: it has been shown that human-imperceptible changes can lead to absurd predictions. In many application areas, including physics, robotics, social sciences and life sciences, this motivates the need for robustness and interpretability, so that deep networks can be trusted in practical applications. Interpretable and robust models can be constructed by incorporating prior knowledge within the model or learning process as an inductive bias, thereby regularizing the model, avoiding overfitting, and making the model easier to understand for scientists who are non-machine-learning experts. Already in the last few years researchers from different fields have proposed various combinations of domain knowledge and machine learning and successfully applied these techniques to various applications.

Our workshop proposal AI for Earth sciences seeks to bring cutting edge geoscientific and planetary challenges to the fore for the machine learning and deep learning communities. We seek machine learning interest from major areas encompassed by Earth sciences which include, atmospheric physics, hydrologic sciences, cryosphere science, oceanography, geology, planetary sciences, space weather, volcanism, seismology, geo-health (i.e. water, land, air pollution, environmental epidemics), biosphere, and biogeosciences. We also seek interest in AI applied to energy for renewable energy meteorology, thermodynamics and heat transfer problems. We call for papers demonstrating novel machine learning techniques in remote sensing for meteorology and geosciences, generative Earth system modeling, and transfer learning from geophysics and numerical simulations and uncertainty in Earth science learning representations. We also seek theoretical developments in interpretable machine learning in meteorology and geoscientific models, hybrid models with Earth science knowledge guided machine learning, representation learning from graphs and manifolds in spatiotemporal models and dimensionality reduction in Earth sciences. In addition, we seek Earth science applications from vision, robotics, multi-agent systems and reinforcement learning. New labelled benchmark datasets and generative visualizations of the Earth are also of particular interest. A new area of interest is in integrated assessment models and human-centered AI for Earth.


AI4Earth Areas of Interest:
- Atmospheric Science
- Hydro and Cryospheres
- Solid Earth
- Theoretical Advances
- Remote Sensing
- Energy in the Earth system
- Extreme weather & climate
- Geo-health
- Biosphere & Biogeosciences
- Planetary sciences
- Benchmark datasets
- People-Earth

Mobile health (mHealth) technologies have transformed the mode and quality of clinical research. Wearable sensors and mobile phones provide real-time data streams that support automated clinical decision making, allowing researchers and clinicians to provide ecological and in-the-moment support to individuals in need. Mobile health technologies are used across various health fields. Their inclusion in clinical care has aimed to improve HIV medication adherence, to increase activity, supplement counseling/pharmacotherapy in treatment for substance use, reinforce abstinence in addictions, and to support recovery from alcohol dependence. The development of mobile health technologies, however, has progressed at a faster pace than the science and methodology to evaluate their validity and efficacy.


Current mHealth technologies are limited in their ability to understand how adverse health behaviors develop, how to predict them, and how to encourage healthy behaviors. In order for mHealth to progress and have expanded impact, the field needs to facilitate collaboration among machine learning researchers, statisticians, mobile sensing researchers, human-computer interaction researchers, and clinicians. Techniques from multiple fields can be brought to bear on the substantive problems facing this interdisciplinary discipline: experimental design, causal inference, multi-modal complex data analytics, representation learning, reinforcement learning, deep learning, transfer learning, data visualization, and clinical integration.

This workshop will assemble researchers from the key areas in this interdisciplinary space necessary to better address the challenges currently facing the widespread use of mobile health technologies.

Communication is one of the most impressive human abilities but historically it has been studied in machine learning mainly on confined datasets of natural language. Thanks to deep RL, emergent communication can now be studied in complex multi-agent scenarios.

Three previous successful workshops (2017-2019) have gathered the community to discuss how, when, and to what end communication emerges, producing research later published at top ML venues (e.g., ICLR, ICML, AAAI). However, many approaches to studying emergent communication rely on extensive amounts of shared training time. Our question is: Can we do that faster?

Humans interact with strangers on a daily basis. They possess a basic shared protocol, but a huge partition is nevertheless defined by the context. Humans are capable of adapting their shared protocol to ever new situations and general AI would need this capability too.

We want to explore the possibilities for artificial agents of evolving ad hoc communication spontaneously, by interacting with strangers. Since humans excel on this task, we want to start by having the participants of the workshop take the role of their agents and develop their own bots for an interactive game. This will illuminate the necessities of zero-shot communication learning in a practical way and form a base of understanding to build algorithms upon. The participants will be split into groups and will have one hour to develop their bots. Then, a round-robin tournament will follow, where bots will play an iterated zero-shot communication game with other teams’ bots.

This interactive approach is especially aimed at the defined NeurIPS workshop goals to clarify questions for a subfield or application area and to crystallize common problems. It condenses our experience from former workshops on how workshop design can facilitate cooperation and progress in the field. We also believe that this will maximize the interactions and exchange of ideas between our community.

https://twitter.com/svrhm2020 The goal of the 2nd Shared Visual Representations in Human and Machine Intelligence (SVRHM) workshop is to disseminate relevant, parallel findings in the fields of computational neuroscience, psychology, and cognitive science that may inform modern machine learning. In the past few years, machine learning methods---especially deep neural networks---have widely permeated the vision science, cognitive science, and neuroscience communities. As a result, scientific modeling in these fields has greatly benefited, producing a swath of potentially critical new insights into the human mind. Since human performance remains the gold standard for many tasks, these cross-disciplinary insights and analytical tools may point towards solutions to many of the current problems that machine learning researchers face (e.g., adversarial attacks, compression, continual learning, and self-supervised learning). Thus we propose to invite leading cognitive scientists with strong computational backgrounds to disseminate their findings to the machine learning community with the hope of closing the loop by nourishing new ideas and creating cross-disciplinary collaborations. In particular, this year's version of the workshop will have a heavy focus on the relative roles of larger datasets and stronger inductive biases as we work on tasks that go beyond object recognition.

Second session for the competition program at NeurIPS2020.

Machine learning competitions have grown in popularity and impact over the last decade, emerging as an effective means to advance the state of the art by posing well-structured, relevant, and challenging problems to the community at large. Motivated by a reward or merely the satisfaction of seeing their machine learning algorithm reach the top of a leaderboard, practitioners innovate, improve, and tune their approach before evaluating on a held-out dataset or environment. The competition track of NeurIPS has matured in 2020, its fourth year, with a considerable increase in both the number of challenges and the diversity of domains and topics. A total of 16 competitions are featured this year as part of the track, with 8 competitions associated to each of the two days. The list of competitions that ar part of the program are available here:

https://neurips.cc/Conferences/2020/CompetitionTrack

Machine learning is rapidly becoming an integral component of sociotechnical systems. Predictions are increasingly used to grant beneficial resources or withhold opportunities, and the consequences of such decisions induce complex social dynamics by changing agent outcomes and prompting individuals to proactively respond to decision rules. This introduces challenges for standard machine learning methodology. Static measurements and training sets poorly capture the complexity of dynamic interactions between algorithms and humans. Strategic adaptation to decision rules can render statistical regularities obsolete. Correlations momentarily observed in data may not be robust enough to support interventions for long-term welfaremits of traditional, static approaches to decision-making, researchers in fields ranging from public policy to computer science to economics have recently begun to view consequential decision-making through a dynamic lens. This workshop will confront the use of machine learning to make consequential decisions in dynamic environments. Work in this area sits at the nexus of several different fields, and the workshop will provide an opportunity to better understand and synthesize social and technical perspectives on these issues and catalyze conversations between researchers and practitioners working across these diverse areas.

Spurred on by recent advances in neural modeling and wet-lab methods, structural biology, the study of the three-dimensional (3D) atomic structure of proteins and other macromolecules, has emerged as an area of great promise for machine learning. The shape of macromolecules is intrinsically linked to their biological function (e.g., much like the shape of a bike is critical to its transportation purposes), and thus machine learning algorithms that can better predict and reason about these shapes promise to unlock new scientific discoveries in human health as well as increase our ability to design novel medicines.

Moreover, fundamental challenges in structural biology motivate the development of new learning systems that can more effectively capture physical inductive biases, respect natural symmetries, and generalize across atomic systems of varying sizes and granularities. Through the Machine Learning in Structural Biology workshop, we aim to include a diverse range of participants and spark a conversation on the required representations and learning algorithms for atomic systems, as well as dive deeply into how to integrate these with novel wet-lab capabilities.

Natural disasters are one of the oldest threats to both individuals and the societies they co-exist in. As a result, humanity has ceaselessly sought way to provide assistance to people in need after disasters have struck. Further, natural disasters are but a single, extreme example of the many possible humanitarian crises. Disease outbreak, famine, and oppression against disadvantaged groups can pose even greater dangers to people that have less obvious solutions. In this proposed workshop, we seek to bring together the Artificial Intelligence (AI) and Humanitarian Assistance and Disaster Response (HADR) communities in order to bring AI to bear on real-world humanitarian crises. Through this workshop, we intend to establish meaningful dialogue between the communities.

By the end of the workshop, the NeurIPS research community can come to understand the practical challenges of aiding those who are experiencing crises, while the HADR community can understand the landscape that is the state of art and practice in AI. Through this, we seek to begin establishing a pipeline of transitioning the research created by the NeurIPS community to real-world humanitarian issues.

Human involvement in AI system design, development, and evaluation is critical to ensure that the insights being derived are practical, and the systems built are meaningful, reliable, and relatable to those who need them. Humans play an integral role in all stages of machine learning development, be it during data generation, interactively teaching machines, or interpreting, evaluating and debugging models. With growing interest in such “human in the loop” learning, we aim to highlight new and emerging research opportunities for the ML community that arise from the evolving needs to design evaluation and training strategies for humans and models in the loop. The specific focus of this workshop is on emerging and under-explored areas of human- and model-in-the-loop learning, such as employing humans to seek richer forms of feedback for data than labels alone, learning from dynamic adversarial data collection with humans employed to find weaknesses in models, learning from human teachers instructing computers through conversation and/or demonstration, investigating the role of humans in model interpretability, and assessing social impact of ML systems. This workshop aims to bring together interdisciplinary researchers from academia and industry to discuss major challenges, outline recent advances, and facilitate future research in these areas.

In the recent decade, we have witnessed rapid progress in machine learning in general and deep learning in particular, mostly driven by tremendous data. As these intelligent algorithms, systems, and applications are deployed in real-world scenarios, we are now facing new challenges, such as scalability, security, privacy, trust, cost, regulation, and environmental and societal impacts. In the meantime, data privacy and ownership has become more and more critical in many domains, such as finance, health, government, and social networks. Federated learning (FL) has emerged to address data privacy issues. To make FL practically scalable, useful, efficient, and effective on security and privacy mechanisms and policies, it calls for joint efforts from the community, academia, and industry. More challenges, interplays, and tradeoffs in scalability, privacy, and security need to be investigated in a more holistic and comprehensive manner by the community. We are expecting broader, deeper, and greater evolution of these concepts and technologies, and confluence towards holistic trustworthy AI ecosystems.

This workshop provides an open forum for researchers, practitioners, and system builders to exchange ideas, discuss, and shape roadmaps towards scalable and privacy-preserving federated learning in particular, and scalable and trustworthy AI ecosystems in general.

Reinforcement Learning (RL) has had numerous successes in recent years in solving complex problem domains. However, this progress has been largely limited to domains where a simulator is available or the real environment is quick and easy to access. This is one of a number of challenges that are bottlenecks to deploying RL agents on real-world systems. Two recent papers identify nine important challenges that, if solved, will take a big step towards enabling RL agents to be deployed to real-world systems (Dulac et. al. 2019, 2020).The goals of this workshop are four-fold: (1) Providing a forum for researchers in academia, industry researchers as well as industry practitioners from diverse backgrounds to discuss the challenges faced in real-world systems; (2) discuss and prioritize the nine research challenges. This includes determining which challenges we should focus on next, whether any new challenges should be added to the list or existing ones removed from this list; (3) Discuss problem formulations for the various challenges and critique these formulations or develop new ones. This is especially important for more abstract challenges such as explainability. We should also be asking ourselves whether the current Markov Decision Process (MDP) formulation is sufficient for solving these problems or whether modifications need to be made. (4) Discuss approaches to solving combinations of these challenges.

There are many tasks that could be automated by writing computer programs, but most people don’t know how to program computers (this is the subject of program synthesis, the study of how to automatically write programs from user specifications). Building tools for doing computer-assisted-programming could thus improve the lives of many people (and it’s also a cool research problem!). There has been substantial recent interest in the ML community in the problem of automatically writing computer programs from user specifications, as evidenced by the increased volume of Program Synthesis submissions to ICML, ICLR, and NeurIPS.

Despite this recent work, a lot of exciting questions are still open, such as how to combine symbolic reasoning over programs with deep learning, how to represent programs and user specifications, and how to apply program synthesis within computer vision, robotics, and other control problems. There is also work to be done on fusing work done in the ML community with research on Programming Languages (PL) through collaboration between the ML and PL communities, and there remains the challenge of establishing benchmarks that allow for easy comparison and measurement of progress. The aim of the CAP workshop is to address these points. This workshop will bring together researchers in programming languages, machine learning, and related areas who are interested in program synthesis and other methods for automatically writing programs from a specification of intended behavior.

Self-supervised learning (SSL) is an unsupervised approach for representation learning without relying on human-provided labels. It creates auxiliary tasks on unlabeled input data and learns representations by solving these tasks. SSL has demonstrated great success on images (e.g., MoCo, PIRL, SimCLR) and texts (e.g., BERT) and has shown promising results in other data modalities, including graphs, time-series, audio, etc. On a wide variety of tasks, SSL without using human-provided labels achieves performance that is close to fully supervised approaches.

The existing SSL research mostly focuses on improving the empirical performance without a theoretical foundation. While the proposed SSL approaches are empirically effective, theoretically why they perform well is not clear. For example, why certain auxiliary tasks in SSL perform better than others? How many unlabeled data examples are needed by SSL to learn a good representation? How is the performance of SSL affected by neural architectures?

In this workshop, we aim to bridge this gap between theory and practice. We bring together SSL-interested researchers from various domains to discuss the theoretical foundations of empirically well-performing SSL approaches and how the theoretical insights can further improve SSL’s empirical performance. Different from previous SSL-related workshops which focus on empirical effectiveness of SSL approaches without considering their theoretical foundations, our workshop focuses on establishing the theoretical foundation of SSL and providing theoretical insights for developing new SSL approaches.
We invite submissions of both theoretical works and empirical works, and the intersection of the two. The topics include but are not limited to:
Theoretical foundations of SSL
Sample complexity of SSL methods
Theory-driven design of auxiliary tasks in SSL
Comparative analysis of different auxiliary tasks
Comparative analysis of SSL and supervised approaches
Information theory and SSL
SSL for computer vision, natural language processing, robotics, speech processing, time-series analysis, graph analytics, etc.
SSL for healthcare, social media, neuroscience, biology, social science, etc.
Cognitive foundations of SSL

In addition to invited talks by leading researchers from diverse backgrounds including CV, NLP, robotics, theoretical ML, etc., the workshop will feature poster sessions and panel discussion to share perspectives on establishing foundational understanding of existing SSL approaches and theoretically-principled ways of developing new SSL methods. We accept submissions of short papers (up to 4 pages excluding references in NeurIPS format), which will be peer-reviewed by at least two reviewers. The accepted papers are allowed to be submitted to other conference venues.

NeurIPS 2020 Workshop on Machine Learning for Systems

Website: http://mlforsystems.org/

Submission Link: https://cmt3.research.microsoft.com/MLFS2020/Submission/Index

Important Dates:

Submission Deadline: October 9th, 2020 (AoE)
Acceptance Notifications: October 23rd, 2020
Camera-Ready Submission: November 29th, 2020
Workshop: December 12th, 2020

Call for Papers:

Machine Learning for Systems is an interdisciplinary workshop that brings together researchers in computer systems and machine learning. This workshop is meant to serve as a platform to promote discussions between researchers in these target areas.

We invite submission of up to 4-page extended abstracts in the broad area of using machine learning in the design of computer systems. We are especially interested in submissions that move beyond using machine learning to replace numerical heuristics. This year, we hope to see novel system designs, streamlined cross-platform optimization, and new benchmarks for ML for Systems.

Accepted papers will be made available on the workshop website, but there will be no formal proceedings. Authors may therefore publish their work in other journals or conferences. The workshop will include invited talks from industry and academia as well as oral and poster presentations by workshop participants.

Areas of interest:

* Supervised, unsupervised, and reinforcement learning research with applications to:
- Systems Software
- Runtime Systems
- Distributed Systems
- Security
- Compilers, data structures, and code optimization
- Databases
- Computer architecture, microarchitecture, and accelerators
- Circuit design and layout
- Interconnects and Networking
- Storage
- Datacenters
* Representation learning for hardware and software
* Optimization of computer systems and software
* Systems modeling and simulation
* Implementations of ML for Systems and challenges
* High quality datasets for ML for Systems problems

Submission Instructions:

We welcome submissions of up to 4 pages (not including references). This is not a strict limit, but authors are encouraged to adhere to it if possible. All submissions must be in PDF format and should follow the NeurIPS 2020 format. Submissions do not have to be anonymized.

Please submit your paper no later than October 9th, 2020 midnight anywhere in the world to CMT (Link available soon).

The common paradigm in reinforcement learning (RL) assumes that an agent frequently interacts with the environment and learns using its own collected experience. This mode of operation is prohibitive for many complex real-world problems, where repeatedly collecting diverse data is expensive (e.g., robotics or educational agents) and/or dangerous (e.g., healthcare). Alternatively, Offline RL focuses on training agents with logged data in an offline fashion with no further environment interaction. Offline RL promises to bring forward a data-driven RL paradigm and carries the potential to scale up end-to-end learning approaches to real-world decision making tasks such as robotics, recommendation systems, dialogue generation, autonomous driving, healthcare systems and safety-critical applications. Recently, successful deep RL algorithms have been adapted to the offline RL setting and demonstrated a potential for success in a number of domains, however, significant algorithmic and practical challenges remain to be addressed. The goal of this workshop is to bring attention to offline RL, both from within and from outside the RL community discuss algorithmic challenges that need to be addressed, discuss potential real-world applications, discuss limitations and challenges, and come up with concrete problem statements and evaluation protocols, inspired from real-world applications, for the research community to work on.

For details on submission please visit: https://offline-rl-neurips.github.io/ (Submission deadline: October 9, 11:59 pm PT)

Speakers:
Emma Brunskill (Stanford)
Finale Doshi-Velez (Harvard)
John Langford (Microsoft Research)
Nan Jiang (UIUC)
Brandyn White (Waymo Research)
Nando de Freitas (DeepMind)

Attempts at understanding deep learning have come from different disciplines, namely physics, statistics, information theory, and machine learning. These lines of investigation have very different modeling assumptions and techniques; it is unclear how their results may be reconciled together. This workshop builds upon the observation that Information Geometry has strong overlaps with these directions and may serve as a means to develop a holistic understanding of deep learning. The workshop program is designed to answer two specific questions. The first question is: how do geometry of the hypothesis class and information-theoretic properties of optimization inform generalization. Good datasets have been a key propeller of the empirical success of deep networks. Our theoretical understanding of data is however poor. The second question the workshop will focus on is: how can we model data and use the understanding of data to improve optimization/generalization in the low-data regime.

Gather.Town link: [ protected link dropped ]