NeurIPS 2023 论文列表

The advancement of Offline Reinforcement Learning (RL) and Offline Multi-Agent Reinforcement Learning (MARL) critically depends on the availability of high-quality, pre-collected offline datasets that represent real-world complexities and practical applications. However, existing datasets often fall short in their simplicity and lack of realism. To address this gap, we propose Hokoff, a comprehensive set of pre-collected datasets that covers both offline RL and offline MARL, accompanied by a robust framework, to facilitate further research. This data is derived from Honor of Kings, a recognized Multiplayer Online Battle Arena (MOBA) game known for its intricate nature, closely resembling real-life situations. Utilizing this framework, we benchmark a variety of offline RL and offline MARL algorithms. We also introduce a novel baseline algorithm tailored for the inherent hierarchical action space of the game. We reveal the incompetency of current offline RL approaches in handling task complexity, generalization and multi-task learning.

There is a growing interest in device-control systems that can interpret human natural language instructions and execute them on a digital device by directly controlling its user interface. We present a dataset for device-control research, Android in the Wild (AitW), which is orders of magnitude larger than current datasets. The dataset contains human demonstrations of device interactions, including the screens and actions, and corresponding natural language instructions. It consists of 715k episodes spanning 30k unique instructions, four versions of Android (v10–13), and eight device types (Pixel 2 XL to Pixel 6) with varying screen resolutions. It contains multi-step tasks that require semantic understanding of language and visual context. This dataset poses a new challenge: actions available through the user interface must be inferred from their visual appearance, and, instead of simple UI element-based actions, the action space consists of precise gestures (e.g., horizontal scrolls to operate carousel widgets). We organize our dataset to encourage robustness analysis of device-control systems, i.e., how well a system performs in the presence of new task descriptions, new applications, or new platform versions. We develop two agents and report performance across the dataset. The dataset is available at https://github.com/google-research/google-research/tree/master/android_in_the_wild.

Historically applied exclusively to perfect information games, depth-limited search with value functions has been key to recent advances in AI for imperfect information games. Most prominent approaches with strong theoretical guarantees require *subgame decomposition* - a process in which a subgame is computed from public information and player beliefs. However, subgame decomposition can itself require non-trivial computations, and its tractability depends on the existence of efficient algorithms for either full enumeration or generation of the histories that form the root of the subgame. Despite this, no formal analysis of the tractability of such computations has been established in prior work, and application domains have often consisted of games, such as poker, for which enumeration is trivial on modern hardware. Applying these ideas to more complex domains requires understanding their cost. In this work, we introduce and analyze the computational aspects and tractability of filtering histories for subgame decomposition. We show that constructing a single history from the root of the subgame is generally intractable, and then provide a necessary and sufficient condition for efficient enumeration. We also introduce a novel Markov Chain Monte Carlo-based generation algorithm for trick-taking card games - a domain where enumeration is often prohibitively expensive. Our experiments demonstrate its improved scalability in the trick-taking card game *Oh Hell*. These contributions clarify when and how depth-limited search via subgame decomposition can be an effective tool for sequential decision-making in imperfect information settings.

As machine learning agents act more autonomously in the world, they will increasingly interact with each other. Unfortunately, in many social dilemmas like the one-shot Prisoner’s Dilemma, standard game theory predicts that ML agents will fail to cooperate with each other. Prior work has shown that one way to enable cooperative outcomes in the one-shot Prisoner’s Dilemma is to make the agents mutually transparent to each other, i.e., to allow them to access one another’s source code (Rubinstein, 1998; Tennenholtz, 2004) – or weights in the case of ML agents. However, full transparency is often unrealistic, whereas partial transparency is commonplace. Moreover, it is challenging for agents to learn their way to cooperation in the full transparency setting. In this paper, we introduce a more realistic setting in which agents only observe a single number indicating how similar they are to each other. We prove that this allows for the same set of cooperative outcomes as the full transparency setting. We also demonstrate experimentally that cooperation can be learned using simple ML methods.

When interacting with people, AI agents do not just influence the state of the world -- they also influence the actions people take in response to the agent, and even their underlying intentions and strategies. Accounting for and leveraging this influence has mostly been studied in settings where it is sufficient to assume that human behavior is near-optimal: competitive games, or general-sum settings like autonomous driving alongside human drivers. Instead, we focus on influence in settings where there is a need to capture human suboptimality. For instance, imagine a collaborative task in which, due either to cognitive biases or lack of information, people do not perform very well -- how could an agent influence them towards more optimal behavior? Assuming near-optimal human behavior will not work here, and so the agent needs to learn from real human data. But experimenting online with humans is potentially unsafe, and creating a high-fidelity simulator of the environment is often impractical. Hence, we focus on learning from an offline dataset of human-human interactions. Our observation is that offline reinforcement learning (RL) can learn to effectively influence suboptimal humans by extending and combining elements of observed human-human behavior. We demonstrate that offline RL can solve two challenges with effective influence. First, we show that by learning from a dataset of suboptimal human-human interaction on a variety of tasks -- none of which contains examples of successful influence -- an agent can learn influence strategies to steer humans towards better performance even on new tasks. Second, we show that by also modeling and conditioning on human behavior, offline RL can learn to affect not just the human's actions but also their underlying strategy, and adapt to changes in their strategy.

Above Ground Biomass is an important variable as forests play a crucial role in mitigating climate change as they act as an efficient, natural and cost-effective carbon sink. Traditional field and airborne LiDAR measurements have been proven to provide reliable estimations of forest biomass. Nevertheless, the use of these techniques at a large scale can be challenging and expensive. Satellite data have been widely used as a valuable tool in estimating biomass on a global scale. However, the full potential of dense multi-modal satellite time series data, in combination with modern deep learning approaches, has yet to be fully explored. The aim of the "BioMassters" data challenge and benchmark dataset is to investigate the potential of multi-modal satellite data (Sentinel-1 SAR and Sentinel-2 MSI) to estimate forest biomass at a large scale using the Finnish Forest Centre's open forest and nature airborne LiDAR data as a reference. The performance of the top three baseline models shows the potential of deep learning to produce accurate and higher-resolution biomass maps. Our benchmark dataset is publically available at https://huggingface.co/datasets/nascetti-a/BioMassters (doi:10.57967/hf/1009) and the implementation of the top three winning models are available at https://github.com/drivendataorg/the-biomassters.

Research on emergent communication between deep-learning-based agents has received extensive attention due to its inspiration for linguistics and artificial intelligence. However, previous attempts have hovered around emerging communication under perception-oriented environmental settings, that forces agents to describe low-level perceptual features intra image or symbol contexts. In this work, inspired by the classic human reasoning test (namely Raven's Progressive Matrix), we propose the Reasoning Game, a cognition-oriented environment that encourages agents to reason and communicate high-level rules, rather than perceived low-level contexts. Moreover, we propose 1) an unbiased dataset (namely rule-RAVEN) as a benchmark to avoid overfitting, 2) and a two-stage curriculum agent training method as a baseline for more stable convergence in the Reasoning Game, where contexts and semantics are bilaterally drifting. Experimental results show that, in the Reasoning Game, a semantically stable and compositional language emerges to solve reasoning problems. The emerged language helps agents apply the extracted rules to the generalization of unseen context attributes, and to the transfer between different context attributes or even tasks.

Influence functions (IF) have been seen as a technique for explaining model predictions through the lens of the training data. Their utility is assumed to be in identifying training examples "responsible" for a prediction so that, for example, correcting a prediction is possible by intervening on those examples (removing or editing them) and retraining the model. However, recent empirical studies have shown that the existing methods of estimating IF predict the leave-one-out-and-retrain effect poorly. In order to understand the mismatch between the theoretical promise and the practical results, we analyse five assumptions made by IF methods which are problematic for modern-scale deep neural networks and which concern convexity, numeric stability, training trajectory and parameter divergence. This allows us to clarify what can be expected theoretically from IF. We show that while most assumptions can be addressed successfully, the parameter divergence poses a clear limitation on the predictive power of IF: influence fades over training time even with deterministic training. We illustrate this theoretical result with BERT and ResNet models. Another conclusion from the theoretical analysis is that IF are still useful for model debugging and correcting even though some of the assumptions made in prior work do not hold: using natural language processing and computer vision tasks, we verify that mis-predictions can be successfully corrected by taking only a few fine-tuning steps on influential examples.

Human Intelligence (HI) excels at combining basic skills to solve complex tasks. This capability is vital for Artificial Intelligence (AI) and should be embedded in comprehensive AI Agents, enabling them to harness expert models for complex task-solving towards Artificial General Intelligence (AGI). Large Language Models (LLMs) show promising learning and reasoning abilities, and can effectively use external models, tools, plugins, or APIs to tackle complex problems. In this work, we introduce OpenAGI, an open-source AGI research and development platform designed for solving multi-step, real-world tasks. Specifically, OpenAGI uses a dual strategy, integrating standard benchmark tasks for benchmarking and evaluation, and open-ended tasks including more expandable models, tools, plugins, or APIs for creative problem-solving. Tasks are presented as natural language queries to the LLM, which then selects and executes appropriate models. We also propose a Reinforcement Learning from Task Feedback (RLTF) mechanism that uses task results to improve the LLM's task-solving ability, which creates a self-improving AI feedback loop. While we acknowledge that AGI is a broad and multifaceted research challenge with no singularly defined solution path, the integration of LLMs with domain-specific expert models, inspired by mirroring the blend of general and specialized intelligence in humans, offers a promising approach towards AGI. We are open-sourcing the OpenAGI project's code, dataset, benchmarks, evaluation methods, and the UI demo to foster community involvement in AGI advancement: https://github.com/agiresearch/OpenAGI.

Image generation techniques have been gaining increasing attention recently, but concerns have been raised about the potential misuse and intellectual property (IP) infringement associated with image generation models. It is, therefore, necessary to analyze the origin of images by inferring if a specific image was generated by a particular model, i.e., origin attribution. Existing methods only focus on specific types of generative models and require additional procedures during the training phase or generation phase. This makes them unsuitable for pre-trained models that lack these specific operations and may impair generation quality. To address this problem, we first develop an alteration-free and model-agnostic origin attribution method via reverse-engineering on image generation models, i.e., inverting the input of a particular model for a specific image. Given a particular model, we first analyze the differences in the hardness of reverse-engineering tasks for generated samples of the given model and other images. Based on our analysis, we then propose a method that utilizes the reconstruction loss of reverse-engineering to infer the origin. Our proposed method effectively distinguishes between generated images of a specific generative model and other images, i.e., images generated by other models and real images.

In this paper, we introduce a generalization of the standard Stackelberg Games (SGs) framework: _Calibrated Stackelberg Games_. In CSGs, a principal repeatedly interacts with an agent who (contrary to standard SGs) does not have direct access to the principal's action but instead best responds to _calibrated forecasts_ about it. CSG is a powerful modeling tool that goes beyond assuming that agents use ad hoc and highly specified algorithms for interacting in strategic settings to infer the principal's actions and thus more robustly addresses real-life applications that SGs were originally intended to capture. Along with CSGs, we also introduce a stronger notion of calibration, termed _adaptive calibration_, that provides fine-grained any-time calibration guarantees against adversarial sequences. We give a general approach for obtaining adaptive calibration algorithms and specialize them for finite CSGs. In our main technical result, we show that in CSGs, the principal can achieve utility that converges to the optimum Stackelberg value of the game both in _finite_ and _continuous_ settings and that no higher utility is achievable. Two prominent and immediate applications of our results are the settings of learning in Stackelberg Security Games and strategic classification, both against _calibrated_ agents.

Algorithmic recommender systems such as Spotify and Netflix affect not only consumer behavior but also *producer incentives*. Producers seek to create content that will be shown by the recommendation algorithm, which can impact both the diversity and quality of their content. In this work, we investigate the resulting supply-side equilibria in personalized content recommender systems. We model the decisions of producers as choosing *multi-dimensional* content vectors and users as having *heterogenous* preferences, which contrasts with classical low-dimensional models. Multi-dimensionality and heterogeneity creates the potential for *specialization*, where different producers create different types of content at equilibrium. Using a duality argument, we derive necessary and sufficient conditions for whether specialization occurs. Then, we characterize the distribution of content at equilibrium in concrete settings with two populations of users. Lastly, we show that specialization can enable producers to achieve *positive profit at equilibrium*, which means that specialization can reduce the competitiveness of the marketplace. At a conceptual level, our analysis of supply-side competition takes a step towards elucidating how personalized recommendations shape the marketplace of digital goods.

In this paper, we introduce a novel approach to fine-grained cross-view geo-localization. Our method aligns a warped ground image with a corresponding GPS-tagged satellite image covering the same area using homography estimation. We first employ a differentiable spherical transform, adhering to geometric principles, to accurately align the perspective of the ground image with the satellite map. This transformation effectively places ground and aerial images in the same view and on the same plane, reducing the task to an image alignment problem. To address challenges such as occlusion, small overlapping range, and seasonal variations, we propose a robust correlation-aware homography estimator to align similar parts of the transformed ground image with the satellite image. Our method achieves sub-pixel resolution and meter-level GPS accuracy by mapping the center point of the transformed ground image to the satellite image using a homography matrix and determining the orientation of the ground camera using a point above the central axis. Operating at a speed of 30 FPS, our method outperforms state-of-the-art techniques, reducing the mean metric localization error by 21.3\% and 32.4\% in same-area and cross-area generalization tasks on the VIGOR benchmark, respectively, and by 34.4\% on the KITTI benchmark in same-area evaluation.

Self-play is a technique for machine learning in multi-agent systems where a learning algorithm learns by interacting with copies of itself. Self-play is useful for generating large quantities of data for learning, but has the drawback that the agents the learner will face post-training may have dramatically different behavior than the learner came to expect by interacting with itself. For the special case of two-player constant-sum games, self-play that reaches Nash equilibrium is guaranteed to produce strategies that perform well against any post-training opponent; however, no such guarantee exists for multiplayer games. We show that in games that approximately decompose into a set of two-player constant-sum games (called constant-sum polymatrix games) where global $\epsilon$-Nash equilibria are boundedly far from Nash equilibria in each subgame (called subgame stability), any no-external-regret algorithm that learns by self-play will produce a strategy with bounded vulnerability. For the first time, our results identify a structural property of multiplayer games that enable performance guarantees for the strategies produced by a broad class of self-play algorithms. We demonstrate our findings through experiments on Leduc poker.

Natural language-conditioned reinforcement learning (RL) enables agents to follow human instructions. Previous approaches generally implemented language-conditioned RL by providing the policy with human instructions in natural language (NL) and training the policy to follow instructions. In this is outside-in approach, the policy must comprehend the NL and manage the task simultaneously. However, the unbounded NL examples often bring much extra complexity for solving concrete RL tasks, which can distract policy learning from completing the task. To ease the learning burden of the policy, we investigate an inside-out scheme for natural language-conditioned RL by developing a task language (TL) that is task-related and easily understood by the policy, thus reducing the policy learning burden. Besides, we employ a translator to translate natural language into the TL, which is used in RL to achieve efficient policy training. We implement this scheme as TALAR (TAsk Language with predicAte Representation) that learns multiple predicates to model object relationships as the TL. Experiments indicate that TALAR not only better comprehends NL instructions but also leads to a better instruction-following policy that significantly improves the success rate over baselines and adapts to unseen expressions of NL instruction. Besides, the TL is also an effective sub-task abstraction compatible with hierarchical RL.

We present a new algorithm, Cross-Episodic Curriculum (CEC), to boost the learning efficiency and generalization of Transformer agents. Central to CEC is the placement of cross-episodic experiences into a Transformer’s context, which forms the basis of a curriculum. By sequentially structuring online learning trials and mixed-quality demonstrations, CEC constructs curricula that encapsulate learning progression and proficiency increase across episodes. Such synergy combined with the potent pattern recognition capabilities of Transformer models delivers a powerful cross-episodic attention mechanism. The effectiveness of CEC is demonstrated under two representative scenarios: one involving multi-task reinforcement learning with discrete control, such as in DeepMind Lab, where the curriculum captures the learning progression in both individual and progressively complex settings; and the other involving imitation learning with mixed-quality data for continuous control, as seen in RoboMimic, where the curriculum captures the improvement in demonstrators' expertise. In all instances, policies resulting from CEC exhibit superior performance and strong generalization. Code is open-sourced on the project website https://cec-agent.github.io/ to facilitate research on Transformer agent learning.

Combining test statistics from independent trials or experiments is a popular method of meta-analysis. However, there is very limited theoretical understanding of the power of the combined test, especially in high-dimensional models considering composite hypotheses tests. We derive a mathematical framework to study standard {meta-analysis} testing approaches in the context of the many normal means model, which serves as the platform to investigate more complex models. We introduce a natural and mild restriction on the meta-level combination functions of the local trials. This allows us to mathematically quantify the cost of compressing $m$ trials into real-valued test statistics and combining these. We then derive minimax lower and matching upper bounds for the separation rates of standard combination methods for e.g. p-values and e-values, quantifying the loss relative to using the full, pooled data. We observe an elbow effect, revealing that in certain cases combining the locally optimal tests in each trial results in a sub-optimal {meta-analysis} method and develop approaches to achieve the global optima. We also explore the possible gains of allowing limited coordination between the trial designs. Our results connect meta-analysis with bandwidth constraint distributed inference and build on recent information theoretic developments in the latter field.

Inspired by the insights in cognitive science with respect to human memory and reasoning mechanism, a novel evolvable LLM-based (Large Language Model) agent framework is proposed as Rememberer. By equipping the LLM with a long-term experience memory, Rememberer is capable of exploiting the experiences from the past episodes even for different task goals, which excels an LLM-based agent with fixed exemplars or equipped with a transient working memory. We further introduce **R**einforcement **L**earning with **E**xperience **M**emory (**RLEM**) to update the memory. Thus, the whole system can learn from the experiences of both success and failure, and evolve its capability without fine-tuning the parameters of the LLM. In this way, the proposed Rememberer constitutes a semi-parametric RL agent. Extensive experiments are conducted on two RL task sets to evaluate the proposed framework. The average results with different initialization and training sets exceed the prior SOTA by 4% and 2% for the success rate on two task sets and demonstrate the superiority and robustness of Rememberer.

Photos serve as a way for humans to record what they experience in their daily lives, and they are often regarded as trustworthy sources of information. However, there is a growing concern that the advancement of artificial intelligence (AI) technology may produce fake photos, which can create confusion and diminish trust in photographs. This study aims to comprehensively evaluate agents for distinguishing state-of-the-art AI-generated visual content. Our study benchmarks both human capability and cutting-edge fake image detection AI algorithms, using a newly collected large-scale fake image dataset Fake2M. In our human perception evaluation, titled HPBench, we discovered that humans struggle significantly to distinguish real photos from AI-generated ones, with a misclassification rate of 38.7\%. Along with this, we conduct the model capability of AI-Generated images detection evaluation MPBench and the top-performing model from MPBench achieves a 13\% failure rate under the same setting used in the human evaluation. We hope that our study can raise awareness of the potential risks of AI-generated images and facilitate further research to prevent the spread of false information. More information can refer to https://github.com/Inf-imagine/Sentry.