Robuta

Code for the paper "Faster Policy Learning with Continuous-Time Gradients" - samuela/ctpg

Nov 19, 2025 - The Bipartisan Learning Communities (BLC) Congressional Fellowship is a no-cost, nonpartisan educational and professional development program for early- to...

The Whole Child Policy Table unifies and coordinates the work of a broad set of policy partners working to create school environments in which every student is...

Growing research demonstrates the benefits for all students of teachers of color. Evidence also points to promising practices to help teachers of color...

The Racial Equity Leadership Network (RELN) is a fellowship program for district leaders who are committed to addressing persistent disparities in their...

Fifty years ago, the Kerner Commission issued a seminal report on racial division and disparities in the United States. This blog series commemorates the...

"A Better Way: Our Vision for a Confident America" doesn't spend a ton of time dealing with K-12, but the plan from a task force of House Republicans and...

After nearly five years of community school implementation, supported by an infusion of state funds, a consortium of small rural districts in California’s...

Nov 9, 2025 - Congress recognized Learning Disabilities Awareness Month, there were major shakeups at the Department of Education, and NCLD hosted its second annual LD...

Learned visuomotor policies are capable of performing increasingly complex manipulation tasks. However, most of these policies are trained on data collected...

07/14/20 - Policy gradient methods have shown success in learning control policies for high-dimensional dynamical systems. Their biggest down...

05/16/19 - We consider the problem of imitation learning from a finite set of expert trajectories, without access to reinforcement signals. T...

Advances in neuroscience, developmental, and learning sciences shed light on the ways that teaching practice must transform to support the whole child and...

Evidence shows that adequate and equitable financing of public schools improves a range of outcomes for all students, yet disparities in school funding...

09/07/21 - Learning to play table tennis is a challenging task for robots, due to the variety of the strokes required. Current advances in de...

More and more states and districts are recognizing that recruiting and retaining teachers of color can help meet their students' needs while also helping...

Founded in 2015, the Learning Policy Institute (LPI) conducts and communicates independent, high-quality research to improve education policy and practice.

The pandemic has exacerbated teacher workforce issues that have persisted for at least a decade. Because of these long-standing conditions, even small changes...

We propose a data-driven automated feature engineering framework Fetch.

LPI’s Center for School and System Redesign supports research, practice initiatives, and policy strategies that can transform educational systems to respond...

Internships About 80 percent of students in the Department of Political Science and Public Policy do at least one internship during their time at Elon. View...

Mar 13, 2025 - Don’t miss the 2022 Spring Ed Tech Advocacy and Policy Fly-Ins scheduled for Wednesday, March 9 (virtual) and Thursday, May 12 (in-person).

California is expanding universal prekindergarten (UPK) to all 4-year-olds and eligible 3-year-olds. Survey data show that local educational agencies have...

Offline safe reinforcement learning (OSRL) involves learning a decision-making policy to maximize rewards from a fixed batch of training data to satisfy...

Olivier Kamanda is a senior technology leader specializing in AI governance, content moderation, and U.S. foreign policy. Former State Department Advisor and...

Join RMIT's Post-Carbon Research Centre in collaboration with Sustainable House Day for a panel conversation exploring the latest policy debates, research...

This paper propose a gradient-based framework to generate contextual symbolic policy for Meta-Reinforcement Learning to improve the generalization ability,...

Electronic proceedings of IJCAI 2024

Abstract page for arXiv paper 2512.03973v1: Guided Flow Policy: Learning from High-Value Actions in Offline Reinforcement Learning

Dec 16, 2025 - Members can join our Child Health Policy Learning Collaborative meetings to hear from experts on timely child health and advanced practice nursing issues.

LPI studied 156 California school districts of significant size in which students of color, as well as White students, consistently achieve at higher levels...

04/22/20 - Deep reinforcement learning (RL) algorithms frequently require prohibitive interaction experience to ensure the quality of learned...

Guidance document setting out Scottish Government expectations for education authorities and their partners when meeting the statutory duties in community...

Community Schools Forward is a partnership of four organizations collaborating with education practitioners, researchers, and leaders across the country to...

03/24/20 - Learning visual similarity requires to learn relations, typically between triplets of images. Albeit triplet approaches being powe...

Curriculum, instruction, and assessment focused on deeper learning develop students’ abilities to think critically and solve complex problems, communicate...

Information for students to learn more about Artifical Intelligence (AI).

Be the first to catch the latest from the Learning Policy Institute. Choose from our three newsletters and get fresh publications, expert blogs, event invites,...

A Machine Learning Approach to Analyze and Support Anticorruption Policy by Elliott Ash, Sergio Galletta and Tommaso Giommoni. Published in volume 17, issue 2,...

Sep 24, 2025 - In July 2024, Ursula von der Leyen's Political Guidelines promised to significantly increase funding for a just transition in the next long-term EU...

Join the discussion on this paper page

Abstract page for arXiv paper 2601.09163: CEI: A Unified Interface for Cross-Embodiment Visuomotor Policy Learning in 3D Space

Baolin Peng, Xiujun Li, Lihong Li, Jianfeng Gao, Asli Celikyilmaz, Sungjin Lee, Kam-Fai Wong. Proceedings of the 2017 Conference on Empirical Methods in...

While high-quality preparation is critical for an effective and stable teacher workforce, only 58% of newly credentialed California teachers in 2020–21 were...

California’s Local Control Funding Formula (LCFF), enacted in 2013, created a simpler and more equitable K-12 education finance system. Research indicates it...

Teacher turnover and shortages hurt student outcomes and well-being. Research identifies a number of evidence-based strategies to recruit and retain a diverse,...

How policy makers can support project-based learning.

The Shared Learning Initiative builds the capacity of California communities to advance more equitable and racially just school systems through partnerships...

Jan 14, 2025 - Please read our privacy policy before viewing our website or requesting more information about our online educational games.

Abstract page for arXiv paper 2403.20250v1: Optimal Policy Learning with Observational Data in Multi-Action Scenarios: Estimation, Risk Preference, and...

Abstract page for arXiv paper 2311.02467v3: Individualized Policy Evaluation and Learning under Clustered Network Interference

Electronic proceedings of IJCAI 2022

What does the data say about teacher working conditions in each state? This interactive map rates each state on how attractive and equitable the teaching...

Dec 8, 2025

Abstract page for arXiv paper 1605.09703: Policy learning for time-bounded reachability in Continuous-Time Markov Decision Processes via doubly-stochastic...

Astronomy & Astrophysics (A&A) is an international journal which publishes papers on all aspects of astronomy and astrophysics

The COVID-19 pandemic exposed many fragilities of our economies and deepened existing inequalities and imbalances. Green fiscal policies in particular can...

Abstract page for arXiv paper 2602.03823: Preference-based Conditional Treatment Effects and Policy Learning

Importance sampling is a central idea underlying off-policy prediction in reinforcement learning. It provides a strategy for re-weighting samples from a...

Humans exhibit diverse and expressive whole-body movements. However, attaining human-like whole-body coordination in humanoid robots remains challenging, as...

Join the discussion on this paper page

AI systems deployed in practical settings (e.g., conversation systems, recommender systems) naturally collect user feedback. Alignment is an important goal of...

Research shows that the use of restorative practices improves achievement, reduces disciplinary and achievement gaps, and supports students’ mental health...

LPI’s 2025 annual scan of teacher shortage data indicates that little has changed over the past year, with many schools continuing to face staffing...

Robust reinforcement learning (RL) aims to find a policy that optimizes the worst-case performance in the face of uncertainties. In this paper, we focus on...

The Learning Policy Institute conducts and communicates independent, high-quality research to improve education policy and practice.

Student engagement with Idaho Policy Institute is built on a foundation of experiential learning for...

Establishes policy for students to receive credit for prior learning.

Hear more from policymakers and academics.

To design reward that align with human goals, Reinforcement Learning from Human Feedback (RLHF) has emerged as a prominent technique for learning reward...

Around the country, severe teacher shortages worsened by the pandemic have put greater pressure on teachers and administrators to scramble to cover positions....

Desiree Carver-Thomas is a Senior Researcher and Policy Analyst at the Learning Policy Institute and co-leads LPI’s Educator Quality team.

The Educator Preparation Laboratory (EdPrepLab) is an initiative developed to support and strengthen high-quality educator preparation in the United States.

A whole child education prioritizes the full scope of a child’s developmental needs to advance equity and ensure that every child reaches their fullest...

To conduct policy efficiently, central banks must use available data to infer, or learn, the relevant structural relationships in the economy. However, because...

This Privacy Policy sets forth the practices in effect at the Southern Poverty Law Center (SPLC) and Learning for Justice to protect the privacy of website...

Effective policy action requires a shared understanding among a wide number of stakeholders of what STI policy is and how innovation emerges and contributes to...

CALVIN - A benchmark for Language-Conditioned Policy Learning for Long-Horizon Robot Manipulation Tasks - mees/calvin

Simulation-based reinforcement learning (RL) has significantly advanced humanoid locomotion tasks, yet direct real-world RL from scratch or starting from...

Aug 25, 2025 - The adoption of Kenya’s 2010 Constitution was celebrated as a beacon of hope for Kenya’s drylands, which the post-colonial centralised power structure had

In reinforcement learning, experience replay stores past samples for further reuse. Prioritized sampling is a promising technique to better utilize these...

Learning visuomotor policy for multi-task robotic manipulation has been a long-standing challenge for the robotics community. The difficulty lies in the...

Building generalist robot policies that can handle diverse tasks in open-ended environments is a central challenge in robotics. To leverage knowledge from...

A severe special education teacher shortage threatens the ability to improve outcomes for students with disabilities, who often have the greatest needs but...