Recent advancements in large language models (LLMs) have resulted in increasingly anthropomorphic language concerning the ability of LLMs to reason. Whether \textit{reasoning} in LLMs should be understood to be inherently different is, however, widely debated. We propose utilizing a representation engineering approach wherein model activations are read from the residual stream of an LLM when processing a reasoning task. The activations are used to derive a control vector that is applied to the model as an inference-time intervention, modulating the representational space of the model, to improve performance on the specified task. We publish the code for deriving control vectors and analyzing model representations. The method allows us to improve performance on reasoning benchmarks and assess how control vectors influence the final logit distribution of a model via metrics such as KL divergence and entropy. We apply control vectors to Mistral-7B-Instruct and a range of Pythia models on an inductive, a deductive and mathematical reasoning task. We show that an LLM can, to a certain degree, be controlled to improve its perceived reasoning ability by modulating activations. The intervention is dependent upon the ability to reliably extract the model's typical state when correctly solving a task. Our results suggest that reasoning performance can be modulated in the same manner as other information-processing tasks performed by LLMs and demonstrate that we are capable of improving performance on specific tasks via a simple intervention on the residual stream with no additional training.

Reinforcement learning (RL) for complex tasks remains a challenge, primarily due to the difficulties of engineering scalar reward functions and the inherent inefficiency of training models from scratch. Instead, it would be better to specify complex tasks in terms of elementary subtasks and to reuse subtask solutions whenever possible. In this work, we address continuous space lexicographic multi-objective RL problems, consisting of prioritized subtasks, which are notoriously difficult to solve. We show that these can be scalarized with a subtask transformation and then solved incrementally using value decomposition. Exploiting this insight, we propose prioritized soft Q-decomposition (PSQD), a novel algorithm for learning and adapting subtask solutions under lexicographic priorities in continuous state-action spaces. PSQD offers the ability to reuse previously learned subtask solutions in a zero-shot composition, followed by an adaptation step. Its ability to use retained subtask training data for offline learning eliminates the need for new environment interaction during adaptation. We demonstrate the efficacy of our approach by presenting successful learning, reuse, and adaptation results for both low- and high-dimensional simulated robot control tasks, as well as offline learning results. In contrast to baseline approaches, PSQD does not trade off between conflicting subtasks or priority constraints and satisfies subtask priorities during learning. PSQD provides an intuitive framework for tackling complex RL problems, offering insights into the inner workings of the subtask composition.

Generative adversarial networks conditioned on textual image descriptions are capable of generating realistic-looking images. However, current methods still struggle to generate images based on complex image captions from a heterogeneous domain. Furthermore, quantitatively evaluating these text-to-image models is challenging, as most evaluation metrics only judge image quality but not the conformity between the image and its caption. To address these challenges we introduce a new model that explicitly models individual objects within an image and a new evaluation metric called Semantic Object Accuracy (SOA) that specifically evaluates images given an image caption. The SOA uses a pre-trained object detector to evaluate if a generated image contains objects that are mentioned in the image caption, e.g., whether an image generated from "a car driving down the street" contains a car. We perform a user study comparing several text-to-image models and show that our SOA metric ranks the models the same way as humans, whereas other metrics such as the Inception Score do not. Our evaluation also shows that models which explicitly model objects outperform models which only model global image characteristics.

Recurrent neural networks that can capture temporal characteristics on multiple timescales are a key architecture in machine learning solutions as well as in neurocognitive models. A crucial open question is how these architectures can adopt both multi-term dependencies and systematic fluctuations from the data or from sensory input, similar to the adaptation and abstraction capabilities of the human brain. In this paper, we propose an extension of the classic Continuous Time Recurrent Neural Network (CTRNN) by allowing it to learn to gate its timescale characteristic during activation and thus dynamically change the timescales in processing sequences. This mechanism is simple but bio-plausible as it is motivated by the modulation of oscillation modes between neural populations. We test how the novel Gating Adaptive CTRNNs can solve difficult synthetic sequence prediction problems and explore the development of the timescale characteristics as well as the interplay of multiple timescales. As a particularly interesting finding, we report that timescale distributions emerge, which simultaneously capture systematic patterns as well as spontaneous fluctuations. Our extended architecture is interesting for cognitive models that aim to investigate the development of specific timescale characteristic under temporally complex perception and action, and vice versa.

In contrast to many established emotion recognition systems, convolutional neural networks do not rely on handcrafted features to categorize emotions. Although achieving state-of-the-art performances, it is still not fully understood what these networks learn and how the learned representations correlate with the emotional characteristics of speech. The aim of this work is to contribute to a deeper understanding of the acoustic and prosodic features that are relevant for the perception of emotional states. Firstly, an artificial deep neural network architecture is proposed that learns the auditory features directly from the raw and unprocessed speech signal. Secondly, we introduce two novel methods for the analysis of the implicitly learned representations based on data-driven and network-driven visualization techniques. Using these methods, we identify how the network categorizes an audio signal as a two-dimensional representation of emotions, namely valence and arousal. The proposed approach is a general method to enable a deeper analysis and understanding of the most relevant representations to perceive emotional expressions in speech.

In this paper, we present an autonomous AI system designed for a Human-Robot Interaction (HRI) study, set around a dice game scenario. We conduct a case study to answer our research question: Does a robot with a socially engaged personality lead to a higher acceptance than a competitive personality? The flexibility of our proposed system allows us to construct and attribute two different personalities to a humanoid robot: a socially engaged personality that maximizes its user interaction and a competitive personality that is focused on playing and winning the game. We evaluate both personalities in a user study, in which the participants play a turn-taking dice game with the robot. Each personality is assessed with four different evaluation tools: 1) the Godspeed Questionnaire, 2) the Mind Perception Questionnaire, 3) a custom questionnaire concerning the overall HRI experience, and 4) a Convolutional Neural Network analyzing the emotions on the participants' facial feedback throughout the game. Our results show that the socially engaged personality evokes stronger emotions among the participants and is rated higher in likability and animacy than the competitive one. We conclude that designing the robot with a socially engaged personality contributes to a higher acceptance within an HRI scenario.

We investigate hierarchical attention networks for the task of question answering. For this purpose, we propose two different approaches: in the first, a document vector representation is built hierarchically from word-to-sentence level which is then used to infer the right answer. In the second, pointer sum attention is utilized to directly infer an answer from the attention values of the word and sentence representations. We evaluate our approach on the Children's Book Test, a cloze-style question answering dataset, and analyze the generated attention distributions. Our results show that, although a hierarchical approach does not offer much improvement over a shallow baseline, it does indeed offer a large performance boost when combining word and sentence attention with pointer sum attention.

Tracking arbitrary objects in natural environments is a challenging task in visual computing. A central problem is the need to adapt to changing appearances under strong transformation and occlusion. We propose a tracking framework that utilises the strength of Convolutional Neural Networks to create a robust and adaptive model of the object from training data produced during tracking. An incremental update mechanism provides increased performance and reduces the computational costs for training during tracking, allowing for robust real-time tracking with state-of-the-art performance. Together with optimisations for deploying the framework on humanoid robots and distributed devices, this shows its viability for research in developmental robotics on questions around infant cognition or active exploration.

For the complex human brain that enables us to communicate in natural language, we gathered good understandings of principles underlying language acquisition and processing, knowledge about sociocultural conditions, and insights into activity patterns in the brain. However, we were not yet able to understand the behavioural and mechanistic characteristics for natural language and how mechanisms in the brain allow to acquire and process language. In bridging the insights from behavioural psychology and neuroscience, the goal of this paper is to contribute a computational understanding of appropriate characteristics that favour language acquisition. Accordingly, we provide concepts and refinements in cognitive modelling regarding principles and mechanisms in the brain and propose a neurocognitively plausible model for embodied language acquisition from real-world interaction of a humanoid robot with its environment. In particular, the architecture consists of a continuous time recurrent neural network, where parts have different leakage characteristics and thus operate on multiple timescales for every modality and the association of the higher level nodes of all modalities into cell assemblies. The model is capable of learning language production grounded in both, temporal dynamic somatosensation and vision, and features hierarchical concept abstraction, concept decomposition, multi-modal integration, and self-organisation of latent representations.

Advancements in Human-Robot Interaction involve robots being more responsive and adaptive to the human user they are interacting with. For example, robots model a personalised dialogue with humans, adapting the conversation to accommodate the user's preferences in order to allow natural interactions. This study investigates the impact of such personalised interaction capabilities of a human companion robot on its social acceptance, perceived intelligence and likeability in a human-robot interaction scenario. In order to measure this impact, the study makes use of an object learning scenario where the user teaches different objects to the robot using natural language. An interaction module is built on top of the learning scenario which engages the user in a personalised conversation before teaching the robot to recognise different objects. The two systems, i.e. with and without the interaction module, are compared with respect to how different users rate the robot on its intelligence and sociability. Although the system equipped with personalised interaction capabilities is rated lower on social acceptance, it is perceived as more intelligent and likeable by the users.

Interdisciplinary research, drawing from robotics, artificial intelligence, neuroscience, psychology, and cognitive science, is a cornerstone to advance the state-of-the-art in multimodal human-robot interaction and neuro-cognitive modeling. Research on neuro-cognitive models benefits from the embodiment of these models into physical, humanoid agents that possess complex, human-like sensorimotor capabilities for multimodal interaction with the real world. For this purpose, we develop and introduce NICO (Neuro-Inspired COmpanion), a humanoid developmental robot that fills a gap between necessary sensing and interaction capabilities and flexible design. This combination makes it a novel neuro-cognitive research platform for embodied sensorimotor computational and cognitive models in the context of multimodal interaction as shown in our results.

This paper describes the techniques used in the submitted video presenting an interaction scenario, realised using the Neuro-Inspired Companion (NICO) robot. NICO engages the users in a personalised conversation where the robot always tracks the users' face, remembers them and interacts with them using natural language. NICO can also learn to perform tasks such as remembering and recalling objects and thus can assist users in their daily chores. The interaction system helps the users to interact as naturally as possible with the robot, enriching their experience with the robot, making it more interesting and engaging.

Automatic speech recognition (ASR) technology has been developed to such a level that off-the-shelf distributed speech recognition services are available (free of cost), which allow researchers to integrate speech into their applications with little development effort or expert knowledge leading to better results compared with previously used open-source tools. Often, however, such services do not accept language models or grammars but process free speech from any domain. While results are very good given the enormous size of the search space, results frequently contain out-of-domain words or constructs that cannot be understood by subsequent domain-dependent natural language understanding (NLU) components. We present a versatile post-processing technique based on phonetic distance that integrates domain knowledge with open-domain ASR results, leading to improved ASR performance. Notably, our technique is able to make use of domain restrictions using various degrees of domain knowledge, ranging from pure vocabulary restrictions via grammars or N-Grams to restrictions of the acceptable utterances. We present results for a variety of corpora (mainly from human-robot interaction) where our combined approach significantly outperforms Google ASR as well as a plain open-source ASR solution.

The early recognition of potentially harmful traffic situations is an important goal of vision-based driver assistance systems. Pedestrians, in particular children, are highly endangered in inner city traffic. Within the DaimlerChrysler urban traffic assistance (UTA) project, we are using stereo vision and motion analysis in order to manage those situations. The flow/depth constraint combines both methods in an elegant way and leads to a robust and powerful detection scheme. A ball bouncing on the road often implies a child crossing the street. Since balls appear very small in the images of our cameras and can move considerably fast, a special algorithm has been developed to achieve maximum recognition reliability.

In this paper we give a closer analysis of the elementary properties of the Banach spaces C(K), where K is a totally disconnected, compact Hausdorff space, in terms of the Boolean algebra B(K) of clopen subsets of K. In particular we sharpen a result in [4] by showing that if B(K1) and B(K2) satisfy the same sentences with ≤ n alternations of quantifiers, then the same is true of C(K1) and C(K2). As a consequence we show that for each n there exist C(K) spaces which are elementarily equivalent for sentences with ≤ n quantifier alternations, but which are not elementary equivalent in the full sense. Thus the elementary properties of Banach spaces cannot be determined by looking at sentences with a bounded number of quantifier alternations. The notion of elementary equivalence for Banach spaces which is studied here was introduced by the second author [4] and is expressed using the language of positive bounded formulas in a first-order language for Banach spaces. As was shown in [4], two Banach spaces are elementarily equivalent in this sense if and only if they have isometrically isomorphic Banach space ultrapowers (or, equivalently, isometrically isomorphic nonstandard hulls.) We consider Banach spaces over the field of real numbers. If X is such a space, Bx will denote the closed unit ball of X, Bx = {x ϵ X∣ ∣∣x∣∣ ≤ 1}. Given a compact Hausdorff space K, we let C(K) denote the Banach space of all continuous real-valued functions on K, under the supremum norm. We will especially be concerned with such spaces when K is a totally disconnected compact Hausdorff space. In that case B(K) will denote the Boolean algebra of all clopen subsets of K. We adopt the standard notation from model theory and Banach space theory.

This paper is a continuation of the authors' paper [7]; in particular, we give a sharper and more useful criterion for the approximate elementary equivalence of Cσ(K) spaces, where K is a totally disconnected compact Hausdorff space. (See Theorem 2 below.) As an application, we obtain a complete description of the Banach spaces X which are approximately equivalent to c0. Namely, X ≡ Ac0 iff X = Cσ(K) where K is a totally disconnected compact Hausdorff space which has a dense set of isolated points and σ is an involutory homeomorphism on K which has a unique fixed point t, and t is not an isolated point. (See Theorem 7.) These results are derived from an analysis which we give of the elementary theories of structures (B, σ), where B is a Boolean algebra and σ is an involutory automorphism of B which leaves at most one nontrivial ultrafilter invariant. Let U(σ) denote this ultrafilter, if it exists; let U(σ) = B in case σ leaves no nontrivial ultrafilter invariant. We show that the elementary theory of (B, σ) is completely determined by the theory of (B, U(σ)) (and conversely, because U(σ) is definable in (B, σ)). This makes possible the use of the explicit invariants given by Éršov [4] for structures (B, U) where U is an ultrafilter on B. (These generalize the Tarski invariants for Boolean algebras [15].) We also use the Éršov invariants in the proof of our main result.