Jaan Aru

Jaan Aru, Matthew Larkum, James M. Shine

Interactions with large language models have led to the suggestion that these models may soon be conscious. From the perspective of neuroscience, this position is difficult to defend. For one, the inputs to large language models lack the embodied, embedded information content characteristic of our sensory contact with the world around us. Secondly, the architecture of large language models is missing key features of the thalamocortical system that have been linked to conscious awareness in mammals. Finally, the evolutionary and developmental trajectories that led to the emergence of living conscious organisms arguably have no parallels in artificial systems as envisioned today. The existence of living organisms depends on their actions, and their survival is intricately linked to multi-level cellular, inter-cellular, and organismal processes culminating in agency and consciousness.

Mrugsen Nagsen Gopnarayan, Jaan Aru, Sebastian Gluth

Over the past decades, cognitive neuroscientists and behavioral economists have recognized the value of describing the process of decision making in detail and modeling the emergence of decisions over time. For example, the time it takes to decide can reveal more about an agent's true hidden preferences than only the decision itself. Similarly, data that track the ongoing decision process such as eye movements or neural recordings contain critical information that can be exploited, even if no decision is made. Here, we argue that artificial intelligence (AI) research would benefit from a stronger focus on insights about how decisions emerge over time and incorporate related process data to improve AI predictions in general and human-AI interactions in particular. First, we introduce a highly established computational framework that assumes decisions to emerge from the noisy accumulation of evidence, and we present related empirical work in psychology, neuroscience, and economics. Next, we discuss to what extent current approaches in multi-agent AI do or do not incorporate process data and models of decision making. Finally, we outline how a more principled inclusion of the evidence-accumulation framework into the training and use of AI can help to improve human-AI interactions in the future.

Jaan Aru, Kristjan-Julius Laak

The vision of a universal AI tutor has remained elusive, despite decades of effort. Could LLMs be the game-changer? We overview novel issues arising from developing a nationwide AI tutor. We highlight the practical questions that point to specific gaps in our scientific understanding of the learning process.

Aqeel Labash, Ardi Tampuu, Tambet Matiisen et al.

Assisted by neural networks, reinforcement learning agents have been able to solve increasingly complex tasks over the last years. The simulation environment in which the agents interact is an essential component in any reinforcement learning problem. The environment simulates the dynamics of the agents' world and hence provides feedback to their actions in terms of state observations and external rewards. To ease the design and simulation of such environments this work introduces $\texttt{APES}$, a highly customizable and open source package in Python to create 2D grid-world environments for reinforcement learning problems. $\texttt{APES}$ equips agents with algorithms to simulate any field of vision, it allows the creation and positioning of items and rewards according to user-defined rules, and supports the interaction of multiple agents.

Jaan Aru

Artificial intelligence (AI) systems capable of generating creative outputs are reshaping our understanding of creativity. This shift presents an opportunity for creativity researchers to reevaluate the key components of the creative process. In particular, the advanced capabilities of AI underscore the importance of studying the internal processes of creativity. This paper explores the neurobiological machinery that underlies these internal processes and describes the experiential component of creativity. It is concluded that although the products of artificial and human creativity can be similar, the internal processes are different. The paper also discusses how AI may negatively affect the internal processes of human creativity, such as the development of skills, the integration of knowledge, and the diversity of ideas.

Andres Campero, Derek Shiller, Jaan Aru et al.

We develop a taxonomical framework for classifying challenges to the possibility of consciousness in digital artificial intelligence systems. This framework allows us to identify the level of granularity at which a given challenge is intended (the levels we propose correspond to Marr's levels) and to disambiguate its degree of force: is it a challenge to computational functionalism that leaves the possibility of digital consciousness open (degree 1), a practical challenge to digital consciousness that suggests improbability without claiming impossibility (degree 2), or an argument claiming that digital consciousness is strictly impossible (degree 3)? We apply this framework to 14 prominent examples from the scientific and philosophical literature. Our aim is not to take a side in the debate, but to provide structure and a tool for disambiguating between challenges to computational functionalism and challenges to digital consciousness, as well as between different ways of parsing such challenges.

Mototaka Suzuki, Jaan Aru

Technological advances in the past decades have begun to enable neuroscientists to address fundamental questions about consciousness in an unprecedented way. Here we review remarkable recent progress in our understanding of cellular-level mechanisms of conscious processing in the brain. Of particular interest are the cortical pyramidal neurons -- or "psychic cells" called by Ramón y Cajal more than 100 years ago -- which have an intriguing cellular mechanism that accounts for selective disruption of feedback signaling in the brain upon anesthetic-induced loss of consciousness. Importantly, a particular class of metabotropic receptors distributed over the dendrites of pyramidal cells are highlighted as the key cellular mechanism. After all, Cajal's instinct over a century ago may turn out to be correct -- we may have just begun to understand whether and how psychic cells indeed generate and control our consciousness.

Brett Puppart, Paul-Henry Paltmann, Jaan Aru

This study introduces "Haunted House" a novel text-based game designed to compare the performance of humans and large language models (LLMs) in model-based reasoning. Players must escape from a house containing nine rooms in a 3x3 grid layout while avoiding the ghost. They are guided by verbal clues that they get each time they move. In Study 1, the results from 98 human participants revealed a success rate of 31.6%, significantly outperforming seven state-of-the-art LLMs tested. Out of 140 attempts across seven LLMs, only one attempt resulted in a pass by Claude 3 Opus. Preliminary results suggested that GPT o3-mini-high performance might be higher, but not at the human level. Further analysis of 29 human participants' moves in Study 2 indicated that LLMs frequently struggled with random and illogical moves, while humans exhibited such errors less frequently. Our findings suggest that current LLMs encounter difficulties in tasks that demand active model-based reasoning, offering inspiration for future benchmarks.

Tarun Khajuria, Braian Olmiro Dias, Marharyta Domnich et al.

In this work, we interpret the representations of multi-object scenes in vision encoders through the lens of structured representations. Structured representations allow modeling of individual objects distinctly and their flexible use based on the task context for both scene-level and object-specific tasks. These capabilities play a central role in human reasoning and generalization, allowing us to abstract away irrelevant details and focus on relevant information in a compact and usable form. We define structured representations as those that adhere to two specific properties: binding specific object information into discrete representation units and segregating object representations into separate sets of tokens to minimize cross-object entanglement. Based on these properties, we evaluated and compared image encoders pre-trained on classification (ViT), large vision-language models (CLIP, BLIP, FLAVA), and self-supervised methods (DINO, DINOv2). We examine the token representations by creating object-decoding tasks that measure the ability of specific tokens to capture individual objects in multi-object scenes from the COCO dataset. This analysis provides insights into how object-wise representations are distributed across tokens and layers within these vision encoders. Our findings highlight significant differences in the representation of objects depending on their relevance to the pre-training objective, with this effect particularly pronounced in the CLS token (often used for downstream tasks). Meanwhile, networks and layers that exhibit more structured representations retain better information about individual objects. To guide practical applications, we propose formal measures to quantify the two properties of structured representations, aiding in selecting and adapting vision encoders for downstream tasks.

Jaan Aru, Aqeel Labash, Oriol Corcoll et al.

Theory of Mind is an essential ability of humans to infer the mental states of others. Here we provide a coherent summary of the potential, current progress, and problems of deep learning approaches to Theory of Mind. We highlight that many current findings can be explained through shortcuts. These shortcuts arise because the tasks used to investigate Theory of Mind in deep learning systems have been too narrow. Thus, we encourage researchers to investigate Theory of Mind in complex open-ended environments. Furthermore, to inspire future deep learning systems we provide a concise overview of prior work done in humans. We further argue that when studying Theory of Mind with deep learning, the research's main focus and contribution ought to be opening up the network's representations. We recommend researchers use tools from the field of interpretability of AI to study the relationship between different network components and aspects of Theory of Mind.

Aqeel Labash, Jaan Aru, Tambet Matiisen et al.

Perspective taking is the ability to take the point of view of another agent. This skill is not unique to humans as it is also displayed by other animals like chimpanzees. It is an essential ability for social interactions, including efficient cooperation, competition, and communication. Here we present our progress toward building artificial agents with such abilities. We implemented a perspective taking task inspired by experiments done with chimpanzees. We show that agents controlled by artificial neural networks can learn via reinforcement learning to pass simple tests that require perspective taking capabilities. We studied whether this ability is more readily learned by agents with information encoded in allocentric or egocentric form for both their visual perception and motor actions. We believe that, in the long run, building better artificial agents with perspective taking ability can help us develop artificial intelligence that is more human-like and easier to communicate with.

Tambet Matiisen, Aqeel Labash, Daniel Majoral et al.

Inferring other agents' mental states such as their knowledge, beliefs and intentions is thought to be essential for effective interactions with other agents. Recently, multiagent systems trained via deep reinforcement learning have been shown to succeed in solving different tasks, but it remains unclear how each agent modeled or represented other agents in their environment. In this work we test whether deep reinforcement learning agents explicitly represent other agents' intentions (their specific aims or goals) during a task in which the agents had to coordinate the covering of different spots in a 2D environment. In particular, we tracked over time the performance of a linear decoder trained to predict the final goal of all agents from the hidden state of each agent's neural network controller. We observed that the hidden layers of agents represented explicit information about other agents' goals, i.e. the target landmark they ended up covering. We also performed a series of experiments, in which some agents were replaced by others with fixed goals, to test the level of generalization of the trained agents. We noticed that during the training phase the agents developed a differential preference for each goal, which hindered generalization. To alleviate the above problem, we propose simple changes to the MADDPG training algorithm which leads to better generalization against unseen agents. We believe that training protocols promoting more active intention reading mechanisms, e.g. by preventing simple symmetry-breaking solutions, is a promising direction towards achieving a more robust generalization in different cooperative and competitive tasks.

Jaan Aru, Raul Vicente

Can deep learning (DL) guide our understanding of computations happening in biological brain? We will first briefly consider how DL has contributed to the research on visual object recognition. In the main part we will assess whether DL could also help us to clarify the computations underlying higher cognitive functions such as Theory of Mind. In addition, we will compare the objectives and learning signals of brains and machines, leading us to conclude that simply scaling up the current DL algorithms will most likely not lead to human level Theory of Mind.

Ardi Tampuu, Tambet Matiisen, Dorian Kodelja et al.

Multiagent systems appear in most social, economical, and political situations. In the present work we extend the Deep Q-Learning Network architecture proposed by Google DeepMind to multiagent environments and investigate how two agents controlled by independent Deep Q-Networks interact in the classic videogame Pong. By manipulating the classical rewarding scheme of Pong we demonstrate how competitive and collaborative behaviors emerge. Competitive agents learn to play and score efficiently. Agents trained under collaborative rewarding schemes find an optimal strategy to keep the ball in the game as long as possible. We also describe the progression from competitive to collaborative behavior. The present work demonstrates that Deep Q-Networks can become a practical tool for studying the decentralized learning of multiagent systems living in highly complex environments.

Madis Vasser, Markus Kängsepp, Jaan Aru

In the present change blindness study subjects explored stereoscopic three dimensional (3D) environments through a virtual reality (VR) headset. A novel method that tracked the subjects' head movements was used for inducing changes in the scene whenever the changing object was out of the field of view. The effect of change location (foreground or background in 3D depth) on change blindness was investigated. Two experiments were conducted, one in the lab (n = 50) and the other online (n = 25). Up to 25% of the changes were undetected and the mean overall search time was 27 seconds in the lab study. Results indicated significantly lower change detection success and more change cycles if the changes occurred in the background, with no differences in overall search times. The results confirm findings from previous studies and extend them to 3D environments. The study also demonstrates the feasibility of online VR experiments.