Henryk Michalewski

Anthony Brohan, Noah Brown, Justice Carbajal et al. · stanford

We study how vision-language models trained on Internet-scale data can be incorporated directly into end-to-end robotic control to boost generalization and enable emergent semantic reasoning. Our goal is to enable a single end-to-end trained model to both learn to map robot observations to actions and enjoy the benefits of large-scale pretraining on language and vision-language data from the web. To this end, we propose to co-fine-tune state-of-the-art vision-language models on both robotic trajectory data and Internet-scale vision-language tasks, such as visual question answering. In contrast to other approaches, we propose a simple, general recipe to achieve this goal: in order to fit both natural language responses and robotic actions into the same format, we express the actions as text tokens and incorporate them directly into the training set of the model in the same way as natural language tokens. We refer to such category of models as vision-language-action models (VLA) and instantiate an example of such a model, which we call RT-2. Our extensive evaluation (6k evaluation trials) shows that our approach leads to performant robotic policies and enables RT-2 to obtain a range of emergent capabilities from Internet-scale training. This includes significantly improved generalization to novel objects, the ability to interpret commands not present in the robot training data (such as placing an object onto a particular number or icon), and the ability to perform rudimentary reasoning in response to user commands (such as picking up the smallest or largest object, or the one closest to another object). We further show that incorporating chain of thought reasoning allows RT-2 to perform multi-stage semantic reasoning, for example figuring out which object to pick up for use as an improvised hammer (a rock), or which type of drink is best suited for someone who is tired (an energy drink).

Aakanksha Chowdhery, Sharan Narang, Jacob Devlin et al. · deepmind, stanford

Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pods. We demonstrate continued benefits of scaling by achieving state-of-the-art few-shot learning results on hundreds of language understanding and generation benchmarks. On a number of these tasks, PaLM 540B achieves breakthrough performance, outperforming the finetuned state-of-the-art on a suite of multi-step reasoning tasks, and outperforming average human performance on the recently released BIG-bench benchmark. A significant number of BIG-bench tasks showed discontinuous improvements from model scale, meaning that performance steeply increased as we scaled to our largest model. PaLM also has strong capabilities in multilingual tasks and source code generation, which we demonstrate on a wide array of benchmarks. We additionally provide a comprehensive analysis on bias and toxicity, and study the extent of training data memorization with respect to model scale. Finally, we discuss the ethical considerations related to large language models and discuss potential mitigation strategies.

Aitor Lewkowycz, Anders Andreassen, David Dohan et al. · deepmind

Language models have achieved remarkable performance on a wide range of tasks that require natural language understanding. Nevertheless, state-of-the-art models have generally struggled with tasks that require quantitative reasoning, such as solving mathematics, science, and engineering problems at the college level. To help close this gap, we introduce Minerva, a large language model pretrained on general natural language data and further trained on technical content. The model achieves state-of-the-art performance on technical benchmarks without the use of external tools. We also evaluate our model on over two hundred undergraduate-level problems in physics, biology, chemistry, economics, and other sciences that require quantitative reasoning, and find that the model can correctly answer nearly a third of them.

David Dohan, Winnie Xu, Aitor Lewkowycz et al. · anthropic, deepmind

Prompted models have demonstrated impressive few-shot learning abilities. Repeated interactions at test-time with a single model, or the composition of multiple models together, further expands capabilities. These compositions are probabilistic models, and may be expressed in the language of graphical models with random variables whose values are complex data types such as strings. Cases with control flow and dynamic structure require techniques from probabilistic programming, which allow implementing disparate model structures and inference strategies in a unified language. We formalize several existing techniques from this perspective, including scratchpads / chain of thought, verifiers, STaR, selection-inference, and tool use. We refer to the resulting programs as language model cascades.

Pengcheng Yin, Wen-Ding Li, Kefan Xiao et al. · stanford

Computational notebooks, such as Jupyter notebooks, are interactive computing environments that are ubiquitous among data scientists to perform data wrangling and analytic tasks. To measure the performance of AI pair programmers that automatically synthesize programs for those tasks given natural language (NL) intents from users, we build ARCADE, a benchmark of 1082 code generation problems using the pandas data analysis framework in data science notebooks. ARCADE features multiple rounds of NL-to-code problems from the same notebook. It requires a model to understand rich multi-modal contexts, such as existing notebook cells and their execution states as well as previous turns of interaction. To establish a strong baseline on this challenging task, we develop PaChiNCo, a 62B code language model (LM) for Python computational notebooks, which significantly outperforms public code LMs. Finally, we explore few-shot prompting strategies to elicit better code with step-by-step decomposition and NL explanation, showing the potential to improve the diversity and explainability of model predictions.

Chrisantha Fernando, Dylan Banarse, Henryk Michalewski et al.

Popular prompt strategies like Chain-of-Thought Prompting can dramatically improve the reasoning abilities of Large Language Models (LLMs) in various domains. However, such hand-crafted prompt-strategies are often sub-optimal. In this paper, we present Promptbreeder, a general-purpose self-referential self-improvement mechanism that evolves and adapts prompts for a given domain. Driven by an LLM, Promptbreeder mutates a population of task-prompts, and subsequently evaluates them for fitness on a training set. Crucially, the mutation of these task-prompts is governed by mutation-prompts that the LLM generates and improves throughout evolution in a self-referential way. That is, Promptbreeder is not just improving task-prompts, but it is also improving the mutationprompts that improve these task-prompts. Promptbreeder outperforms state-of-the-art prompt strategies such as Chain-of-Thought and Plan-and-Solve Prompting on commonly used arithmetic and commonsense reasoning benchmarks. Furthermore, Promptbreeder is able to evolve intricate task-prompts for the challenging problem of hate speech classification.

Kuang-Huei Lee, Ofir Nachum, Mengjiao Yang et al.

A longstanding goal of the field of AI is a method for learning a highly capable, generalist agent from diverse experience. In the subfields of vision and language, this was largely achieved by scaling up transformer-based models and training them on large, diverse datasets. Motivated by this progress, we investigate whether the same strategy can be used to produce generalist reinforcement learning agents. Specifically, we show that a single transformer-based model - with a single set of weights - trained purely offline can play a suite of up to 46 Atari games simultaneously at close-to-human performance. When trained and evaluated appropriately, we find that the same trends observed in language and vision hold, including scaling of performance with model size and rapid adaptation to new games via fine-tuning. We compare several approaches in this multi-game setting, such as online and offline RL methods and behavioral cloning, and find that our Multi-Game Decision Transformer models offer the best scalability and performance. We release the pre-trained models and code to encourage further research in this direction.

Szymon Tworkowski, Konrad Staniszewski, Mikołaj Pacek et al.

Large language models have an exceptional capability to incorporate new information in a contextual manner. However, the full potential of such an approach is often restrained due to a limitation in the effective context length. One solution to this issue is to endow an attention layer with access to an external memory, which comprises of (key, value) pairs. Yet, as the number of documents increases, the proportion of relevant keys to irrelevant ones decreases, leading the model to focus more on the irrelevant keys. We identify a significant challenge, dubbed the distraction issue, where keys linked to different semantic values might overlap, making them hard to distinguish. To tackle this problem, we introduce the Focused Transformer (FoT), a technique that employs a training process inspired by contrastive learning. This novel approach enhances the structure of the (key, value) space, enabling an extension of the context length. Our method allows for fine-tuning pre-existing, large-scale models to lengthen their effective context. This is demonstrated by our fine-tuning of $3B$ and $7B$ OpenLLaMA checkpoints. The resulting models, which we name LongLLaMA, exhibit advancements in tasks requiring a long context. We further illustrate that our LongLLaMA models adeptly manage a $256 k$ context length for passkey retrieval.

Spyridon Mouselinos, Mateusz Malinowski, Henryk Michalewski

Recently, high-performing code generation systems based on large language models have surfaced. They are trained on massive corpora containing much more natural text than actual executable computer code. This work shows that current code generation systems exhibit undesired biases inherited from their large language model backbones, which can reduce the quality of the generated code under specific circumstances. To investigate the effect, we propose the "block of influence" concept, which enables a modular decomposition and analysis of the coding challenges. We introduce an automated intervention mechanism reminiscent of adversarial testing that exposes undesired biases through the failure modes of the models under test. Finally, we demonstrate how our framework can be used as a data transformation technique during fine-tuning, acting as a mitigation strategy for these biases.

Thang Luong, Dawsen Hwang, Hoang H. Nguyen et al.

Finding the right north-star metrics is highly critical for advancing the mathematical reasoning capabilities of foundation models, especially given that existing evaluations are either too easy or only focus on getting correct short answers. To address these issues, we present IMO-Bench, a suite of advanced reasoning benchmarks, vetted by a panel of top specialists and that specifically targets the level of the International Mathematical Olympiad (IMO), the most prestigious venue for young mathematicians. IMO-AnswerBench first tests models on 400 diverse Olympiad problems with verifiable short answers. IMO-Proof Bench is the next-level evaluation for proof-writing capabilities, which includes both basic and advanced IMO level problems as well as detailed grading guidelines to facilitate automatic grading. These benchmarks played a crucial role in our historic achievement of the gold-level performance at IMO 2025 with Gemini Deep Think (Luong and Lockhart, 2025). Our model achieved 80.0% on IMO-AnswerBench and 65.7% on the advanced IMO-Proof Bench, surpassing the best non-Gemini models by large margins of 6.9% and 42.4% respectively. We also showed that autograders built with Gemini reasoning correlate well with human evaluations and construct IMO-GradingBench, with 1000 human gradings on proofs, to enable further progress in automatic evaluation of long-form answers. We hope that IMO-Bench will help the community towards advancing robust mathematical reasoning and release it at https://imobench.github.io/.

George Tsoukalas, Anton Kovsharov, Sergey Shirobokov et al.

Large language models (LLMs) increasingly excel at mathematical reasoning, but their unreliability limits their utility in mathematics research. A mitigation is using LLMs to generate formal proofs in languages like Lean. We perform the first large-scale evaluation of this method's ability to solve open problems. Our most capable agent autonomously resolved 9 of 353 open Erdős problems at the per-problem cost of a few hundred dollars, proved 44/492 OEIS conjectures, and is being deployed in combinatorics, optimization, graph theory, algebraic geometry, and quantum optics research. A basic agent alternating LLM-based generation with Lean-based verification replicated the Erdős successes but proved costlier on the hardest problems. These findings demonstrate the power of AI-aided formal proof search and shed light on the agent designs that enable it.

Gemini Team, Petko Georgiev, Ving Ian Lei et al. · deepmind, mila

In this report, we introduce the Gemini 1.5 family of models, representing the next generation of highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio. The family includes two new models: (1) an updated Gemini 1.5 Pro, which exceeds the February version on the great majority of capabilities and benchmarks; (2) Gemini 1.5 Flash, a more lightweight variant designed for efficiency with minimal regression in quality. Gemini 1.5 models achieve near-perfect recall on long-context retrieval tasks across modalities, improve the state-of-the-art in long-document QA, long-video QA and long-context ASR, and match or surpass Gemini 1.0 Ultra's state-of-the-art performance across a broad set of benchmarks. Studying the limits of Gemini 1.5's long-context ability, we find continued improvement in next-token prediction and near-perfect retrieval (>99%) up to at least 10M tokens, a generational leap over existing models such as Claude 3.0 (200k) and GPT-4 Turbo (128k). Finally, we highlight real-world use cases, such as Gemini 1.5 collaborating with professionals on completing their tasks achieving 26 to 75% time savings across 10 different job categories, as well as surprising new capabilities of large language models at the frontier; when given a grammar manual for Kalamang, a language with fewer than 200 speakers worldwide, the model learns to translate English to Kalamang at a similar level to a person who learned from the same content.

Gheorghe Comanici, Eric Bieber, Mike Schaekermann et al. · amazon-science, baidu

In this report, we introduce the Gemini 2.X model family: Gemini 2.5 Pro and Gemini 2.5 Flash, as well as our earlier Gemini 2.0 Flash and Flash-Lite models. Gemini 2.5 Pro is our most capable model yet, achieving SoTA performance on frontier coding and reasoning benchmarks. In addition to its incredible coding and reasoning skills, Gemini 2.5 Pro is a thinking model that excels at multimodal understanding and it is now able to process up to 3 hours of video content. Its unique combination of long context, multimodal and reasoning capabilities can be combined to unlock new agentic workflows. Gemini 2.5 Flash provides excellent reasoning abilities at a fraction of the compute and latency requirements and Gemini 2.0 Flash and Flash-Lite provide high performance at low latency and cost. Taken together, the Gemini 2.X model generation spans the full Pareto frontier of model capability vs cost, allowing users to explore the boundaries of what is possible with complex agentic problem solving.

Gemma Team, Thomas Mesnard, Cassidy Hardin et al. · deepmind

This work introduces Gemma, a family of lightweight, state-of-the art open models built from the research and technology used to create Gemini models. Gemma models demonstrate strong performance across academic benchmarks for language understanding, reasoning, and safety. We release two sizes of models (2 billion and 7 billion parameters), and provide both pretrained and fine-tuned checkpoints. Gemma outperforms similarly sized open models on 11 out of 18 text-based tasks, and we present comprehensive evaluations of safety and responsibility aspects of the models, alongside a detailed description of model development. We believe the responsible release of LLMs is critical for improving the safety of frontier models, and for enabling the next wave of LLM innovations.

Błażej Osiński, Piotr Miłoś, Adam Jakubowski et al.

This work introduces interactive traffic scenarios in the CARLA simulator, which are based on real-world traffic. We concentrate on tactical tasks lasting several seconds, which are especially challenging for current control methods. The CARLA Real Traffic Scenarios (CRTS) is intended to be a training and testing ground for autonomous driving systems. To this end, we open-source the code under a permissive license and present a set of baseline policies. CRTS combines the realism of traffic scenarios and the flexibility of simulation. We use it to train agents using a reinforcement learning algorithm. We show how to obtain competitive polices and evaluate experimentally how observation types and reward schemes affect the training process and the resulting agent's behavior.

Yeming Wen, Pengcheng Yin, Kensen Shi et al.

Large language models (LLMs) have recently demonstrated a remarkable ability to generate code from natural language (NL) prompts. However, in the real world, NL is often too ambiguous to capture the true intent behind programming problems, requiring additional input-output (I/O) specifications. Unfortunately, LLMs can have difficulty aligning their outputs with both the NL prompt and the I/O specification. In this paper, we give a way to mitigate this issue in the context of data science programming, where tasks require explicit I/O specifications for clarity. Specifically, we propose GIFT4Code, a novel approach for the instruction fine-tuning of LLMs with respect to I/O specifications. Our method leverages synthetic data produced by the LLM itself and utilizes execution-derived feedback as a key learning signal. This feedback, in the form of program I/O specifications, is provided to the LLM to facilitate instruction fine-tuning. We evaluated our approach on two challenging data science benchmarks, Arcade and DS-1000. The results demonstrate a significant improvement in the LLM's ability to generate code that is not only executable but also accurately aligned with user specifications, substantially improving the quality of code generation for complex data science tasks.

Konrad Staniszewski, Szymon Tworkowski, Sebastian Jaszczur et al.

Recent advancements in long-context large language models have attracted significant attention, yet their practical applications often suffer from suboptimal context utilization. This study investigates structuring training data to enhance semantic interdependence, demonstrating that this approach effectively improves context utilization. To this end, we introduce the Structured Packing for Long Context (SPLiCe) method, which utilizes retrieval to collate mutually relevant documents into long and coherent training examples. We validate SPLiCe empirically across models of varying sizes -- 3B, 7B, and 13B -- achieving improved performance in long-context tasks, such as Qasper and HotpotQA. Remarkably, even brief fine-tuning with SPLiCe is sufficient to realize these benefits. Additionally, SPLiCe effectively mitigates the lost-in-middle phenomenon often observed in large models. Our comprehensive analysis of SPLiCe explores its design choices and reveals intriguing transfer effects; for instance, training on programming code enhances performance on natural language tasks.

Spyridon Mouselinos, Henryk Michalewski, Mateusz Malinowski

Large Language Models (LLMs) demonstrate ever-increasing abilities in mathematical and algorithmic tasks, yet their geometric reasoning skills are underexplored. We investigate LLMs' abilities in constructive geometric problem-solving one of the most fundamental steps in the development of human mathematical reasoning. Our work reveals notable challenges that the state-of-the-art LLMs face in this domain despite many successes in similar areas. LLMs exhibit biases in target variable selection and struggle with 2D spatial relationships, often misrepresenting and hallucinating objects and their placements. To this end, we introduce a framework that formulates an LLMs-based multi-agents system that enhances their existing reasoning potential by conducting an internal dialogue. This work underscores LLMs' current limitations in geometric reasoning and improves geometric reasoning capabilities through self-correction, collaboration, and diverse role specializations.

Gemini Team, Rohan Anil, Sebastian Borgeaud et al.

This report introduces a new family of multimodal models, Gemini, that exhibit remarkable capabilities across image, audio, video, and text understanding. The Gemini family consists of Ultra, Pro, and Nano sizes, suitable for applications ranging from complex reasoning tasks to on-device memory-constrained use-cases. Evaluation on a broad range of benchmarks shows that our most-capable Gemini Ultra model advances the state of the art in 30 of 32 of these benchmarks - notably being the first model to achieve human-expert performance on the well-studied exam benchmark MMLU, and improving the state of the art in every one of the 20 multimodal benchmarks we examined. We believe that the new capabilities of the Gemini family in cross-modal reasoning and language understanding will enable a wide variety of use cases. We discuss our approach toward post-training and deploying Gemini models responsibly to users through services including Gemini, Gemini Advanced, Google AI Studio, and Cloud Vertex AI.

Spyridon Mouselinos, Henryk Michalewski, Mateusz Malinowski

How can we measure the reasoning capabilities of intelligence systems? Visual question answering provides a convenient framework for testing the model's abilities by interrogating the model through questions about the scene. However, despite scores of various visual QA datasets and architectures, which sometimes yield even a super-human performance, the question of whether those architectures can actually reason remains open to debate. To answer this, we extend the visual question answering framework and propose the following behavioral test in the form of a two-player game. We consider black-box neural models of CLEVR. These models are trained on a diagnostic dataset benchmarking reasoning. Next, we train an adversarial player that re-configures the scene to fool the CLEVR model. We show that CLEVR models, which otherwise could perform at a human level, can easily be fooled by our agent. Our results put in doubt whether data-driven approaches can do reasoning without exploiting the numerous biases that are often present in those datasets. Finally, we also propose a controlled experiment measuring the efficiency of such models to learn and perform reasoning.

Maxwell Nye, Anders Johan Andreassen, Guy Gur-Ari et al.

Large pre-trained language models perform remarkably well on tasks that can be done "in one pass", such as generating realistic text or synthesizing computer programs. However, they struggle with tasks that require unbounded multi-step computation, such as adding integers or executing programs. Surprisingly, we find that these same models are able to perform complex multi-step computations -- even in the few-shot regime -- when asked to perform the operation "step by step", showing the results of intermediate computations. In particular, we train transformers to perform multi-step computations by asking them to emit intermediate computation steps into a "scratchpad". On a series of increasingly complex tasks ranging from long addition to the execution of arbitrary programs, we show that scratchpads dramatically improve the ability of language models to perform multi-step computations.

Sebastian Jaszczur, Aakanksha Chowdhery, Afroz Mohiuddin et al.

Large Transformer models yield impressive results on many tasks, but are expensive to train, or even fine-tune, and so slow at decoding that their use and study becomes out of reach. We address this problem by leveraging sparsity. We study sparse variants for all layers in the Transformer and propose Scaling Transformers, a family of next generation Transformer models that use sparse layers to scale efficiently and perform unbatched decoding much faster than the standard Transformer as we scale up the model size. Surprisingly, the sparse layers are enough to obtain the same perplexity as the standard Transformer with the same number of parameters. We also integrate with prior sparsity approaches to attention and enable fast inference on long sequences even with limited memory. This results in performance competitive to the state-of-the-art on long text summarization.

Michał Zawalski, Błażej Osiński, Henryk Michalewski et al.

Multi-agent reinforcement learning (MARL) provides a framework for problems involving multiple interacting agents. Despite apparent similarity to the single-agent case, multi-agent problems are often harder to train and analyze theoretically. In this work, we propose MA-Trace, a new on-policy actor-critic algorithm, which extends V-Trace to the MARL setting. The key advantage of our algorithm is its high scalability in a multi-worker setting. To this end, MA-Trace utilizes importance sampling as an off-policy correction method, which allows distributing the computations with no impact on the quality of training. Furthermore, our algorithm is theoretically grounded - we prove a fixed-point theorem that guarantees convergence. We evaluate the algorithm extensively on the StarCraft Multi-Agent Challenge, a standard benchmark for multi-agent algorithms. MA-Trace achieves high performance on all its tasks and exceeds state-of-the-art results on some of them.

Piotr Nawrot, Szymon Tworkowski, Michał Tyrolski et al.

Transformer models yield impressive results on many NLP and sequence modeling tasks. Remarkably, Transformers can handle long sequences which allows them to produce long coherent outputs: full paragraphs produced by GPT-3 or well-structured images produced by DALL-E. These large language models are impressive but also very inefficient and costly, which limits their applications and accessibility. We postulate that having an explicit hierarchical architecture is the key to Transformers that efficiently handle long sequences. To verify this claim, we first study different ways to downsample and upsample activations in Transformers so as to make them hierarchical. We use the best performing upsampling and downsampling layers to create Hourglass - a hierarchical Transformer language model. Hourglass improves upon the Transformer baseline given the same amount of computation and can yield the same results as Transformers more efficiently. In particular, Hourglass sets new state-of-the-art for Transformer models on the ImageNet32 generation task and improves language modeling efficiency on the widely studied enwik8 benchmark.

Jacob Austin, Augustus Odena, Maxwell Nye et al.

This paper explores the limits of the current generation of large language models for program synthesis in general purpose programming languages. We evaluate a collection of such models (with between 244M and 137B parameters) on two new benchmarks, MBPP and MathQA-Python, in both the few-shot and fine-tuning regimes. Our benchmarks are designed to measure the ability of these models to synthesize short Python programs from natural language descriptions. The Mostly Basic Programming Problems (MBPP) dataset contains 974 programming tasks, designed to be solvable by entry-level programmers. The MathQA-Python dataset, a Python version of the MathQA benchmark, contains 23914 problems that evaluate the ability of the models to synthesize code from more complex text. On both datasets, we find that synthesis performance scales log-linearly with model size. Our largest models, even without finetuning on a code dataset, can synthesize solutions to 59.6 percent of the problems from MBPP using few-shot learning with a well-designed prompt. Fine-tuning on a held-out portion of the dataset improves performance by about 10 percentage points across most model sizes. On the MathQA-Python dataset, the largest fine-tuned model achieves 83.8 percent accuracy. Going further, we study the model's ability to engage in dialog about code, incorporating human feedback to improve its solutions. We find that natural language feedback from a human halves the error rate compared to the model's initial prediction. Additionally, we conduct an error analysis to shed light on where these models fall short and what types of programs are most difficult to generate. Finally, we explore the semantic grounding of these models by fine-tuning them to predict the results of program execution. We find that even our best models are generally unable to predict the output of a program given a specific input.

Piotr Piękos, Henryk Michalewski, Mateusz Malinowski

Imagine you are in a supermarket. You have two bananas in your basket and want to buy four apples. How many fruits do you have in total? This seemingly straightforward question can be challenging for data-driven language models, even if trained at scale. However, we would expect such generic language models to possess some mathematical abilities in addition to typical linguistic competence. Towards this goal, we investigate if a commonly used language model, BERT, possesses such mathematical abilities and, if so, to what degree. For that, we fine-tune BERT on a popular dataset for word math problems, AQuA-RAT, and conduct several tests to understand learned representations better. Since we teach models trained on natural language to do formal mathematics, we hypothesize that such models would benefit from training on semi-formal steps that explain how math results are derived. To better accommodate such training, we also propose new pretext tasks for learning mathematical rules. We call them (Neighbor) Reasoning Order Prediction (ROP or NROP). With this new model, we achieve significantly better outcomes than data-driven baselines and even on-par with more tailored models. We also show how to reduce positional bias in such models.

Piotr Kozakowski, Łukasz Kaiser, Henryk Michalewski et al.

Sample efficiency and performance in the offline setting have emerged as significant challenges of deep reinforcement learning. We introduce Q-Value Weighted Regression (QWR), a simple RL algorithm that excels in these aspects. QWR is an extension of Advantage Weighted Regression (AWR), an off-policy actor-critic algorithm that performs very well on continuous control tasks, also in the offline setting, but has low sample efficiency and struggles with high-dimensional observation spaces. We perform an analysis of AWR that explains its shortcomings and use these insights to motivate QWR. We show experimentally that QWR matches the state-of-the-art algorithms both on tasks with continuous and discrete actions. In particular, QWR yields results on par with SAC on the MuJoCo suite and - with the same set of hyperparameters - yields results on par with a highly tuned Rainbow implementation on a set of Atari games. We also verify that QWR performs well in the offline RL setting.

Sebastian Jaszczur, Michał Łuszczyk, Henryk Michalewski

We use neural graph networks with a message-passing architecture and an attention mechanism to enhance the branching heuristic in two SAT-solving algorithms. We report improvements of learned neural heuristics compared with two standard human-designed heuristics.

Błażej Osiński, Adam Jakubowski, Piotr Miłoś et al.

We use reinforcement learning in simulation to obtain a driving system controlling a full-size real-world vehicle. The driving policy takes RGB images from a single camera and their semantic segmentation as input. We use mostly synthetic data, with labelled real-world data appearing only in the training of the segmentation network. Using reinforcement learning in simulation and synthetic data is motivated by lowering costs and engineering effort. In real-world experiments we confirm that we achieved successful sim-to-real policy transfer. Based on the extensive evaluation, we analyze how design decisions about perception, control, and training impact the real-world performance.

Zsolt Zombori, Adrián Csiszárik, Henryk Michalewski et al.

We present a reinforcement learning (RL) based guidance system for automated theorem proving geared towards Finding Longer Proofs (FLoP). Unlike most learning based approaches, we focus on generalising from very little training data and achieving near complete confidence. We use several simple, structured datasets with very long proofs to show that FLoP can successfully generalise a single training proof to a large class of related problems. On these benchmarks, FLoP is competitive with strong theorem provers despite using very limited search, due to its ability to solve problems that are prohibitively long for other systems.

Lukasz Kaiser, Mohammad Babaeizadeh, Piotr Milos et al.

Model-free reinforcement learning (RL) can be used to learn effective policies for complex tasks, such as Atari games, even from image observations. However, this typically requires very large amounts of interaction -- substantially more, in fact, than a human would need to learn the same games. How can people learn so quickly? Part of the answer may be that people can learn how the game works and predict which actions will lead to desirable outcomes. In this paper, we explore how video prediction models can similarly enable agents to solve Atari games with fewer interactions than model-free methods. We describe Simulated Policy Learning (SimPLe), a complete model-based deep RL algorithm based on video prediction models and present a comparison of several model architectures, including a novel architecture that yields the best results in our setting. Our experiments evaluate SimPLe on a range of Atari games in low data regime of 100k interactions between the agent and the environment, which corresponds to two hours of real-time play. In most games SimPLe outperforms state-of-the-art model-free algorithms, in some games by over an order of magnitude.

Michał Garmulewicz, Henryk Michalewski, Piotr Miłoś

We propose an expert-augmented actor-critic algorithm, which we evaluate on two environments with sparse rewards: Montezumas Revenge and a demanding maze from the ViZDoom suite. In the case of Montezumas Revenge, an agent trained with our method achieves very good results consistently scoring above 27,000 points (in many experiments beating the first world). With an appropriate choice of hyperparameters, our algorithm surpasses the performance of the expert data. In a number of experiments, we have observed an unreported bug in Montezumas Revenge which allowed the agent to score more than 800,000 points.

Cezary Kaliszyk, Josef Urban, Henryk Michalewski et al.

We introduce a theorem proving algorithm that uses practically no domain heuristics for guiding its connection-style proof search. Instead, it runs many Monte-Carlo simulations guided by reinforcement learning from previous proof attempts. We produce several versions of the prover, parameterized by different learning and guiding algorithms. The strongest version of the system is trained on a large corpus of mathematical problems and evaluated on previously unseen problems. The trained system solves within the same number of inferences over 40% more problems than a baseline prover, which is an unusually high improvement in this hard AI domain. To our knowledge this is the first time reinforcement learning has been convincingly applied to solving general mathematical problems on a large scale.

Łukasz Kidziński, Sharada Prasanna Mohanty, Carmichael Ong et al.

In the NIPS 2017 Learning to Run challenge, participants were tasked with building a controller for a musculoskeletal model to make it run as fast as possible through an obstacle course. Top participants were invited to describe their algorithms. In this work, we present eight solutions that used deep reinforcement learning approaches, based on algorithms such as Deep Deterministic Policy Gradient, Proximal Policy Optimization, and Trust Region Policy Optimization. Many solutions use similar relaxations and heuristics, such as reward shaping, frame skipping, discretization of the action space, symmetry, and policy blending. However, each of the eight teams implemented different modifications of the known algorithms.

Igor Adamski, Robert Adamski, Tomasz Grel et al.

We present a study in Distributed Deep Reinforcement Learning (DDRL) focused on scalability of a state-of-the-art Deep Reinforcement Learning algorithm known as Batch Asynchronous Advantage ActorCritic (BA3C). We show that using the Adam optimization algorithm with a batch size of up to 2048 is a viable choice for carrying out large scale machine learning computations. This, combined with careful reexamination of the optimizer's hyperparameters, using synchronous training on the node level (while keeping the local, single node part of the algorithm asynchronous) and minimizing the memory footprint of the model, allowed us to achieve linear scaling for up to 64 CPU nodes. This corresponds to a training time of 21 minutes on 768 CPU cores, as opposed to 10 hours when using a single node with 24 cores achieved by a baseline single-node implementation.

Robert Adamski, Tomasz Grel, Maciej Klimek et al.

The asynchronous nature of the state-of-the-art reinforcement learning algorithms such as the Asynchronous Advantage Actor-Critic algorithm, makes them exceptionally suitable for CPU computations. However, given the fact that deep reinforcement learning often deals with interpreting visual information, a large part of the train and inference time is spent performing convolutions. In this work we present our results on learning strategies in Atari games using a Convolutional Neural Network, the Math Kernel Library and TensorFlow 0.11rc0 machine learning framework. We also analyze effects of asynchronous computations on the convergence of reinforcement learning algorithms.

Jakub Sygnowski, Henryk Michalewski

We train a number of neural networks to play games Bowling, Breakout and Seaquest using information stored in the memory of a video game console Atari 2600. We consider four models of neural networks which differ in size and architecture: two networks which use only information contained in the RAM and two mixed networks which use both information in the RAM and information from the screen. As the benchmark we used the convolutional model proposed in NIPS and received comparable results in all considered games. Quite surprisingly, in the case of Seaquest we were able to train RAM-only agents which behave better than the benchmark screen-only agent. Mixing screen and RAM did not lead to an improved performance comparing to screen-only and RAM-only agents.