A fundamental part of human cognition is the ability to not only recall previous memories, but also reason across them to draw conclusions. In cognitive science and psychology, this is termed relational reasoning and a number of effects and biases have been observed in human cognition. Designing experiments to measure these reasoning effects is effortful and does not transfer easily to analyzing language model reasoning patterns. To make exploring language models on relational reasoning easier, we introduce ReCogLab – a generative framework for constructing reasoning examples. Unlike static datasets, our framework has a number of benefits that help us in our goal of flexible evaluation of LLMs. First, our framework allows us to control the difficulty and context-length of the problem, allowing us to scale with model capability and evaluate LLMs at a variety of scales. Second, the ability to change the configuration of a dataset dynamically allows us to probe models on different aspects and capabilities. Finally, the flexibility of our approach enables the recreation of classic cognitive science experiments and the systematic study of relational reasoning biases in language models. We demonstrate several such experiments and present our findings on a wide variety of open and closed-source language models. We release all data and code at https://github.com/google-deepmind/recoglab.

In day-to-day communication, people often approximate the truth — for example, rounding the time or omitting details — in order to be maximally helpful to the listener. How do large language models (LLMs) handle such nuanced trade-offs? To address this question, we use psychological models and experiments designed to characterize human behavior to analyze LLMs. We test a range of LLMs and explore how optimization for human preferences or inference-time reasoning affects these trade-offs. We find that reinforcement learning from human feedback improves both honesty and helpfulness, while chain-of-thought prompting skews LLMs towards helpfulness over honesty. Finally, GPT-4 Turbo demonstrates human-like response patterns including sensitivity to the conversational framing and listener’s decision context. Our findings reveal the conversational values internalized by LLMs and suggest that even these abstract values can, to a degree, be steered by zero-shot prompting.

Vision language models (VLMs) have shown impressive capabilities across a variety of tasks, from logical reasoning to visual understanding. This opens the door to richer interaction with the world, for example robotic control. However, VLMs produce only textual outputs, while robotic control and other spatial tasks require outputting continuous coordinates, actions, or trajectories. How can we enable VLMs to handle such settings without fine-tuning on task-specific data? In this paper, we propose a novel visual prompting approach for VLMs that we call Prompting with Iterative Visual Optimization (PIVOT), which casts tasks as iterative visual question answering. In each iteration, the image is annotated with a visual representation of proposals that the VLM can refer to (e.g., candidate robot actions, localizations, or trajectories). The VLM then selects the best ones for the task. These proposals are iteratively refined, allowing the VLM to eventually zero in on the best available answer. We investigate PIVOT on real-world robotic navigation, real-world manipulation from images, instruction following in simulation, and additional spatial inference tasks such as localization. We find, perhaps surprisingly, that our approach enables zero-shot control of robotic systems without any robot training data, navigation in a variety of environments, and other capabilities. Although current performance is far from perfect, our work highlights potentials and limitations of this new regime and shows a promising approach for Internet-Scale VLMs in robotic and spatial reasoning domains.

Tiwalayo Eisape, Michael Tessler, Ishita Dasgupta, Fei Sha, Sjoerd Steenkiste, Tal Linzen. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.

Jackson Petty, Sjoerd Steenkiste, Ishita Dasgupta, Fei Sha, Dan Garrette, Tal Linzen. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.

With the continuous advancement of large language models (LLMs), it is essential to create new benchmarks to evaluate their expanding capabilities and identify areas for improvement. This work focuses on multi-image reasoning, an emerging capability in state-of-the-art LLMs. We introduce ReMI, a dataset designed to assess LLMs' ability to reason with multiple images. This dataset encompasses a diverse range of tasks, spanning various reasoning domains such as math, physics, logic, code, table/chart understanding, and spatial and temporal reasoning. It also covers a broad spectrum of characteristics found in multi-image reasoning scenarios. We have benchmarked several cutting-edge LLMs using ReMI and found a substantial gap between their performance and human-level proficiency. This highlights the challenges in multi-image reasoning and the need for further research. Our analysis also reveals the strengths and weaknesses of different models, shedding light on the types of reasoning that are currently attainable and areas where future models require improvement. We anticipate that ReMI will be a valuable resource for developing and evaluating more sophisticated LLMs capable of handling real-world multi-image understanding tasks.

Instruction-following agents must ground language into their observation and action spaces. Learning to ground language is challenging, typically requiring domain-specific engineering or large quantities of human interaction data. To address this challenge, we propose using pretrained vision-language models (VLMs) to supervise embodied agents. We combine ideas from model distillation and hindsight experience replay (HER), using a VLM to retroactively generate language describing the agent’s behavior. Simple prompting allows us to control the supervision signal, teaching an agent to interact with novel objects based on their names (e.g., planes) or their features (e.g., colors) in a 3D rendered environment. Fewshot prompting lets us teach abstract category membership, including pre-existing categories (food vs toys) and ad-hoc ones (arbitrary preferences over objects). Our work outlines a new and effective way to use internet-scale VLMs, repurposing the generic language grounding acquired by such models to teach task-relevant groundings to embodied agents.

What can be learned about causality and experimentation from passive data? This question is salient given recent successes of passively-trained language models in interactive domains such as tool use. Passive learning is inherently limited. However, we show that purely passive learning can in fact allow an agent to learn generalizable strategies for determining and using causal structures, as long as the agent can intervene at test time. We formally illustrate that learning a strategy of first experimenting, then seeking goals, can allow generalization from passive learning in principle. We then show empirically that agents trained via imitation on expert data can indeed generalize at test time to infer and use causal links which are never present in the training data; these agents can also generalize experimentation strategies to novel variable sets never observed in training. We then show that strategies for causal intervention and exploitation can be generalized from passive data even in a more complex environment with high-dimensional observations, with the support of natural language explanations. Explanations can even allow passive learners to generalize out-of-distribution from perfectly-confounded training data. Finally, we show that language models, trained only on passive next-word prediction, can generalize causal intervention strategies from a few-shot prompt containing explanations and reasoning. These results highlight the surprising power of passive learning of active causal strategies, and have implications for understanding the behaviors and capabilities of language models.

The ability to acquire abstract knowledge is a hallmark of human intelligence and is believed by many to be one of the core differences between humans and neural network models. Agents can be endowed with an inductive bias towards abstraction through meta-learning, where they are trained on a distribution of tasks that share some abstract structure that can be learned and applied. However, because neural networks are hard to interpret, it can be difficult to tell whether agents have learned the underlying abstraction, or alternatively statistical patterns that are characteristic of that abstraction. In this work, we compare the performance of humans and agents in a meta-reinforcement learning paradigm in which tasks are generated from abstract rules. We define a novel methodology for building "task metamers" that closely match the statistics of the abstract tasks but use a different underlying generative process, and evaluate performance on both abstract and metamer tasks. We find that humans perform better at abstract tasks than metamer tasks whereas common neural network architectures typically perform worse on the abstract tasks than the matched metamers. This work provides a foundation for characterizing differences between humans and machine learning that can be used in future work towards developing machines with more human-like behavior.

Machine learning systems often do not share the same inductive biases as humans and, as a result, extrapolate or generalize in ways that are inconsistent with our expectations. The trade-off between exemplar- and rule-based generalization has been studied extensively in cognitive psychology; in this work, we present a protocol inspired by these experimental approaches to probe the inductive biases that control this trade-off in category-learning systems such as artificial neural networks. We isolate two such inductive biases: feature-level bias (differences in which features are more readily learned) and exemplar-vs-rule bias (differences in how these learned features are used for generalization of category labels). We find that standard neural network models are feature-biased and have a propensity towards exemplar-based extrapolation; we discuss the implications of these findings for machine-learning research on data augmentation, fairness, and systematic generalization.

Inferring the abstract relational and causal structure of the world is a major challenge for reinforcement-learning (RL) agents. For humans, language{—}particularly in the form of explanations{—}plays a considerable role in overcoming this challenge. Here, we show that language can play a similar role for deep RL agents in complex environments. While agents typically struggle to acquire relational and causal knowledge, augmenting their experience by training them to predict language descriptions and explanations can overcome these limitations. We show that language can help agents learn challenging relational tasks, and examine which aspects of language contribute to its benefits. We then show that explanations can help agents to infer not only relational but also causal structure. Language can shape the way that agents to generalize out-of-distribution from ambiguous, causally-confounded training, and explanations even allow agents to learn to perform experimental interventions to identify causal relationships. Our results suggest that language description and explanation may be powerful tools for improving agent learning and generalization.

Strong inductive biases give humans the ability to quickly learn to perform a variety of tasks. Although meta-learning is a method to endow neural networks with useful inductive biases, agents trained by meta-learning may sometimes acquire very different strategies from humans. We show that co-training these agents on predicting representations from natural language task descriptions and programs induced to generate such tasks guides them toward more human-like inductive biases. Human-generated language descriptions and program induction models that add new learned primitives both contain abstract concepts that can compress description length. Co-training on these representations result in more human-like behavior in downstream meta-reinforcement learning agents than less abstract controls (synthetic language descriptions, program induction without learned primitives), suggesting that the abstraction supported by these representations is key.

From learning to play the piano to speaking a new language, reusing and recombining previously acquired representations enables us to master complex skills and easily adapt to new environments. Inspired by the Gestalt principle of \textit{grouping by proximity} and theories of chunking in cognitive science, we propose a hierarchical chunking model (HCM). HCM learns representations from non-i.i.d. sequential data from the ground up by first discovering the minimal atomic sequential units as chunks. As learning progresses, a hierarchy of chunk representations is acquired by chunking previously learned representations into more complex representations guided by sequential dependence. We provide learning guarantees on an idealized version of HCM, and demonstrate that HCM learns meaningful and interpretable representations in a human-like fashion. Our model can be extended to learn visual, temporal, and visual-temporal chunks. The interpretability of the learned chunks can be used to assess transfer or interference when the environment changes. Finally, in an fMRI dataset, we demonstrate that HCM learns interpretable chunks of functional coactivation regions and hierarchical modular and sub-modular structures confirmed by the neuroscientific literature. Taken together, our results show how cognitive science in general and theories of chunking in particular can inform novel and more interpretable approaches to representation learning.

A fundamental ability of an intelligent web-based agent is seeking out and acquiring new information. Internet search engines reliably find the correct vicinity but the top results may be a few links away from the desired target. A complementary approach is navigation via hyperlinks, employing a policy that comprehends local content and selects a link that moves it closer to the target. In this paper, we show that behavioral cloning of randomly sampled trajectories is sufficient to learn an effective link selection policy. We demonstrate the approach on a graph version of Wikipedia with 38M nodes and 387M edges. The model is able to efficiently navigate between nodes 5 and 20 steps apart 96% and 92% of the time, respectively. We then use the resulting embeddings and policy in downstream fact verification and question answering tasks where, in combination with basic TF-IDF search and ranking methods, they are competitive results to the state-of-the-art methods.

Developments in machine learning interpretability techniques over the past decade have provided new tools to observe the image regions that are most informative for classification and localization in artificial neural networks (ANNs). Are the same regions similarly informative to human observers? Using data from 79 new experiments and 7,810 participants, we show that passive attention techniques reveal a significant overlap with human visual selectivity estimates derived from 6 distinct behavioral tasks including visual discrimination, spatial localization, recognizability, free-viewing, cued-object search, and saliency search fixations. We find that input visualizations derived from relatively simple ANN architectures probed using guided backpropagation methods are the best predictors of a shared component in the joint variability of the human measures. We validate these correlational results with causal manipulations using recognition experiments. We show that images masked with ANN attention maps were easier for humans to classify than control masks in a speeded recognition experiment. Similarly, we find that recognition performance in the same ANN models was likewise influenced by masking input images using human visual selectivity maps. This work contributes a new approach to evaluating the biological and psychological validity of leading ANNs as models of human vision: by examining their similarities and differences in terms of their visual selectivity to the information contained in images.