Search

Annotating 3D medical images demands substantial time and expertise, driving the adoption of semi-supervised learning (SSL) for segmentation tasks. However, the complex anatomical structures of organs often lead to significant class imbalances, posing major challenges for deploying SSL in real-world scenarios. Despite the availability of valuable prior information, such as inter-organ relative positions and organ shape priors, existing SSL methods have yet to fully leverage these insights. To address this gap, we propose a novel approach that integrates textual anatomical knowledge (TAK) into the segmentation model. Specifically, we use GPT-4o to generate textual descriptions of anatomical priors, which are then encoded using a CLIP-based model. These encoded priors are injected into the segmentation model as parameters of the segmentation head. Additionally, contrastive learning is employed to enhance the alignment between textual priors and visual features. Extensive experiments demonstrate the superior performance of our method, significantly surpassing state-of-the-art approaches. The source code will be available at: https://github.com/Lunn88/TAK-Semi.

Combining reinforcement learning with language grounding is challenging as the agent needs to explore the environment while simultaneously learning multiple language-conditioned tasks. To address this, we introduce a novel method: the compositionally-enabled reinforcement learning language agent (CERLLA). Our method reduces the sample complexity of tasks specified with language by leveraging compositional policy representations and a semantic parser trained using reinforcement learning and in-context learning. We evaluate our approach in an environment requiring function approximation and demonstrate compositional generalization to novel tasks. Our method significantly outperforms the previous best non-compositional baseline in terms of sample complexity on 162 tasks designed to test compositional generalization. Our model attains a higher success rate and learns in fewer steps than the non-compositional baseline. It reaches a success rate equal to an oracle policy's upper-bound performance of 92%. With the same number of environment steps, the baseline only reaches a success rate of 80%., Comment: TMLR 2024

Modern e-commerce platforms offer vast product selections, making it difficult for customers to find items that they like and that are relevant to their current session intent. This is why it is key for e-commerce platforms to have near real-time scalable and adaptable personalized ranking and search systems. While numerous methods exist in the scientific literature for building such systems, many are unsuitable for large-scale industrial use due to complexity and performance limitations. Consequently, industrial ranking systems often resort to computationally efficient yet simplistic retrieval or candidate generation approaches, which overlook near real-time and heterogeneous customer signals, which results in a less personalized and relevant experience. Moreover, related customer experiences are served by completely different systems, which increases complexity, maintenance, and inconsistent experiences. In this paper, we present a personalized, adaptable near real-time ranking platform that is reusable across various use cases, such as browsing and search, and that is able to cater to millions of items and customers under heavy load (thousands of requests per second). We employ transformer-based models through different ranking layers which can learn complex behavior patterns directly from customer action sequences while being able to incorporate temporal (e.g. in-session) and contextual information. We validate our system through a series of comprehensive offline and online real-world experiments at a large online e-commerce platform, and we demonstrate its superiority when compared to existing systems, both in terms of customer experience as well as in net revenue. Finally, we share the lessons learned from building a comprehensive, modern ranking platform for use in a large-scale e-commerce environment.

Two rods made from sol-gel silica have been doped with Ce ions or co-doped with Ce and Tb ions respectively. First, a thermoluminescence (TL) characterization of the trapping and luminescence parameters is carried out to understand the physical mechanisms involved following the irradiation of these rods and to evaluate their dosimetric properties. The optically stimulated luminescence (OSL) and the radioluminescence (RL) responses as a function of respectively absorbed dose and dose rate are assessed. The OSL response of both rods is linear as a function of the absorbed dose. The RL sensitivity of both rods proves not only to be extremely high but also shows a linear behavior over more than 6 decades, allowing real-time detection of dose rates as low as few $\mu$Gy s$^{-1}$, a threshold that to our knowledge, has never been reached. The results from RL and OSL show that these silica-based doped rods are potentially suitable for medical dosimetry and environmental monitoring around nuclear sites.

This paper summarizes the experiments and results of the HYBRINFOX team for the CheckThat! 2024 - Task 1 competition. We propose an approach enriching Language Models such as RoBERTa with embeddings produced by triples (subject ; predicate ; object) extracted from the text sentences. Our analysis of the developmental data shows that this method improves the performance of Language Models alone. On the evaluation data, its best performance was in English, where it achieved an F1 score of 71.1 and ranked 12th out of 27 candidates. On the other languages (Dutch and Arabic), it obtained more mixed results. Future research tracks are identified toward adapting this processing pipeline to more recent Large Language Models., Comment: Paper to appear in the Proceedings of the Conference and Labs of the Evaluation Forum (CLEF 2024 CheckThat!)

This paper presents the HYBRINFOX method used to solve Task 2 of Subjectivity detection of the CLEF 2024 CheckThat! competition. The specificity of the method is to use a hybrid system, combining a RoBERTa model, fine-tuned for subjectivity detection, a frozen sentence-BERT (sBERT) model to capture semantics, and several scores calculated by the English version of the expert system VAGO, developed independently of this task to measure vagueness and subjectivity in texts based on the lexicon. In English, the HYBRINFOX method ranked 1st with a macro F1 score of 0.7442 on the evaluation data. For the other languages, the method used a translation step into English, producing more mixed results (ranking 1st in Multilingual and 2nd in Italian over the baseline, but under the baseline in Bulgarian, German, and Arabic). We explain the principles of our hybrid approach, and outline ways in which the method could be improved for other languages besides English., Comment: To appear in the Proceedings of the Conference and Labs of the Evaluation Forum (CLEF 2024 CheckThat!)

Reinforcement learning (RL) has progressed substantially over the past decade, with much of this progress being driven by benchmarks. Many benchmarks are focused on video or board games, and a large number of robotics benchmarks lack diversity and real-world applicability. In this paper, we aim to simplify the process of applying reinforcement learning in the 3D simulation league of Robocup, a robotic football competition. To this end, we introduce a Robocup-based RL environment based on the open source rcssserver3d soccer server, simple pre-defined tasks, and integration with a popular RL library, Stable Baselines 3. Our environment enables the creation of high-dimensional continuous control tasks within a robotics football simulation. In each task, an RL agent controls a simulated Nao robot, and can interact with the ball or other agents. We open-source our environment and training code at https://github.com/Michael-Beukman/RobocupGym.

Transformers have revolutionized machine learning across diverse domains, yet understanding their behavior remains crucial, particularly in high-stakes applications. This paper introduces the contextual counting task, a novel toy problem aimed at enhancing our understanding of Transformers in quantitative and scientific contexts. This task requires precise localization and computation within datasets, akin to object detection or region-based scientific analysis. We present theoretical and empirical analysis using both causal and non-causal Transformer architectures, investigating the influence of various positional encodings on performance and interpretability. In particular, we find that causal attention is much better suited for the task, and that no positional embeddings lead to the best accuracy, though rotary embeddings are competitive and easier to train. We also show that out of distribution performance is tightly linked to which tokens it uses as a bias term.

Tomography of single-particle-resolved detectors is of primary importance for characterizing particle correlations with applications in quantum metrology, quantum simulation and quantum computing. However, it is a non-trivial task in practice due to the unavoidable presence of noise that affects the measurement but does not originate from the detector. In this work, we address this problem for a three-dimensional single-atom-resolved detector where shot-to-shot atom number fluctuations are a central issue to perform a quantum detector tomography. We overcome this difficulty by exploiting the parallel measurement of counting statistics in sub-volumes of the detector, from which we evaluate the effect of shot-to-shot fluctuations and perform a local tomography of the detector. In addition, we illustrate the validity of our method from applying it to Gaussian quantum states with different number statistics. Finally, we show that the response of Micro-Channel Plate detectors is well-described from using a binomial distribution with the detection efficiency as a single parameter., Comment: 7 pages, 5 figures

We propose a method to remotely verify the authenticity of Optically Variable Devices (OVDs), often referred to as ``holograms'', in identity documents. Our method processes video clips captured with smartphones under common lighting conditions, and is evaluated on two public datasets: MIDV-HOLO and MIDV-2020. Thanks to a weakly-supervised training, we optimize a feature extraction and decision pipeline which achieves a new leading performance on MIDV-HOLO, while maintaining a high recall on documents from MIDV-2020 used as attack samples. It is also the first method, to date, to effectively address the photo replacement attack task, and can be trained on either genuine samples, attack samples, or both for increased performance. By enabling to verify OVD shapes and dynamics with very little supervision, this work opens the way towards the use of massive amounts of unlabeled data to build robust remote identity document verification systems on commodity smartphones. Code is available at https://github.com/EPITAResearchLab/pouliquen.24.icdar, Comment: Accepted at the International Conference on Document Analysis and Recognition (ICDAR 2024)

We present a corpus of 100 documents, OBSINFOX, selected from 17 sources of French press considered unreliable by expert agencies, annotated using 11 labels by 8 annotators. By collecting more labels than usual, by more annotators than is typically done, we can identify features that humans consider as characteristic of fake news, and compare them to the predictions of automated classifiers. We present a topic and genre analysis using Gate Cloud, indicative of the prevalence of satire-like text in the corpus. We then use the subjectivity analyzer VAGO, and a neural version of it, to clarify the link between ascriptions of the label Subjective and ascriptions of the label Fake News. The annotated dataset is available online at the following url: https://github.com/obs-info/obsinfox Keywords: Fake News, Multi-Labels, Subjectivity, Vagueness, Detail, Opinion, Exaggeration, French Press, Comment: Paper to appear in the Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)

This paper investigates the language of propaganda and its stylistic features. It presents the PPN dataset, standing for Propagandist Pseudo-News, a multisource, multilingual, multimodal dataset composed of news articles extracted from websites identified as propaganda sources by expert agencies. A limited sample from this set was randomly mixed with papers from the regular French press, and their URL masked, to conduct an annotation-experiment by humans, using 11 distinct labels. The results show that human annotators were able to reliably discriminate between the two types of press across each of the labels. We propose different NLP techniques to identify the cues used by the annotators, and to compare them with machine classification. They include the analyzer VAGO to measure discourse vagueness and subjectivity, a TF-IDF to serve as a baseline, and four different classifiers: two RoBERTa-based models, CATS using syntax, and one XGBoost combining syntactic and semantic features., Comment: Paper to appear in the EACL 2024 Proceedings of the Third Workshop on Understanding Implicit and Underspecified Language (UnImplicit 2024)

We present counting reward automata-a finite state machine variant capable of modelling any reward function expressible as a formal language. Unlike previous approaches, which are limited to the expression of tasks as regular languages, our framework allows for tasks described by unrestricted grammars. We prove that an agent equipped with such an abstract machine is able to solve a larger set of tasks than those utilising current approaches. We show that this increase in expressive power does not come at the cost of increased automaton complexity. A selection of learning algorithms are presented which exploit automaton structure to improve sample efficiency. We show that the state machines required in our formulation can be specified from natural language task descriptions using large language models. Empirical results demonstrate that our method outperforms competing approaches in terms of sample efficiency, automaton complexity, and task completion., Comment: 14 pages, 11 Figures, Published in AAAI W25: Neuro-Symbolic Learning and Reasoning in the era of Large Language Models (NuCLeaR)

We present AstroCLIP, a single, versatile model that can embed both galaxy images and spectra into a shared, physically meaningful latent space. These embeddings can then be used - without any model fine-tuning - for a variety of downstream tasks including (1) accurate in-modality and cross-modality semantic similarity search, (2) photometric redshift estimation, (3) galaxy property estimation from both images and spectra, and (4) morphology classification. Our approach to implementing AstroCLIP consists of two parts. First, we embed galaxy images and spectra separately by pretraining separate transformer-based image and spectrum encoders in self-supervised settings. We then align the encoders using a contrastive loss. We apply our method to spectra from the Dark Energy Spectroscopic Instrument and images from its corresponding Legacy Imaging Survey. Overall, we find remarkable performance on all downstream tasks, even relative to supervised baselines. For example, for a task like photometric redshift prediction, we find similar performance to a specifically-trained ResNet18, and for additional tasks like physical property estimation (stellar mass, age, metallicity, and sSFR), we beat this supervised baseline by 19\% in terms of $R^2$. We also compare our results to a state-of-the-art self-supervised single-modal model for galaxy images, and find that our approach outperforms this benchmark by roughly a factor of two on photometric redshift estimation and physical property prediction in terms of $R^2$, while remaining roughly in-line in terms of morphology classification. Ultimately, our approach represents the first cross-modal self-supervised model for galaxies, and the first self-supervised transformer-based architectures for galaxy images and spectra., Comment: 18 pages, accepted in Monthly Notices of the Royal Astronomical Society, Presented at the NeurIPS 2023 AI4Science Workshop

We introduce multiple physics pretraining (MPP), an autoregressive task-agnostic pretraining approach for physical surrogate modeling of spatiotemporal systems with transformers. In MPP, rather than training one model on a specific physical system, we train a backbone model to predict the dynamics of multiple heterogeneous physical systems simultaneously in order to learn features that are broadly useful across systems and facilitate transfer. In order to learn effectively in this setting, we introduce a shared embedding and normalization strategy that projects the fields of multiple systems into a shared embedding space. We validate the efficacy of our approach on both pretraining and downstream tasks over a broad fluid mechanics-oriented benchmark. We show that a single MPP-pretrained transformer is able to match or outperform task-specific baselines on all pretraining sub-tasks without the need for finetuning. For downstream tasks, we demonstrate that finetuning MPP-trained models results in more accurate predictions across multiple time-steps on systems with previously unseen physical components or higher dimensional systems compared to training from scratch or finetuning pretrained video foundation models. We open-source our code and model weights trained at multiple scales for reproducibility.

Due in part to their discontinuous and discrete default encodings for numbers, Large Language Models (LLMs) have not yet been commonly used to process numerically-dense scientific datasets. Rendering datasets as text, however, could help aggregate diverse and multi-modal scientific data into a single training corpus, thereby potentially facilitating the development of foundation models for science. In this work, we introduce xVal, a strategy for continuously tokenizing numbers within language models that results in a more appropriate inductive bias for scientific applications. By training specially-modified language models from scratch on a variety of scientific datasets formatted as text, we find that xVal generally outperforms other common numerical tokenization strategies on metrics including out-of-distribution generalization and computational efficiency., Comment: 15 pages, 12 figures. Appendix: 8 pages, 2 figures. Accepted contribution at the NeurIPS Workshop on ML for the Physical Sciences

An important problem in reinforcement learning is designing agents that learn to solve tasks safely in an environment. A common solution is for a human expert to define either a penalty in the reward function or a cost to be minimised when reaching unsafe states. However, this is non-trivial, since too small a penalty may lead to agents that reach unsafe states, while too large a penalty increases the time to convergence. Additionally, the difficulty in designing reward or cost functions can increase with the complexity of the problem. Hence, for a given environment with a given set of unsafe states, we are interested in finding the upper bound of rewards at unsafe states whose optimal policies minimise the probability of reaching those unsafe states, irrespective of task rewards. We refer to this exact upper bound as the "Minmax penalty", and show that it can be obtained by taking into account both the controllability and diameter of an environment. We provide a simple practical model-free algorithm for an agent to learn this Minmax penalty while learning the task policy, and demonstrate that using it leads to agents that learn safe policies in high-dimensional continuous control environments.

This paper provides a detailed explanation of the steps taken to extract and repair a Google Home's internal data. Starting with reverse engineering the hardware of a commercial off-the-shelf Google Home, internal data is then extracted by desoldering and dumping the flash memory. As error correction is performed by the CPU using an undisclosed method, a new alternative method is shown to repair a corrupted SquashFS filesystem, under the assumption of a single or double bitflip per gzip-compressed fragment. Finally, a new method to handle multiple possible repairs using three-valued logic is presented., Comment: 28 pages, modified version of paper that appeared originally at Forensic Science International: Digital Investigation

We investigate the optical properties of silicon-vacancy (SiV) and germanium-vacancy (GeV) color centers in nanodiamonds under hydrostatic pressure up to 180 GPa. The nanodiamonds were synthetized by Si or Ge-doped plasma assisted chemical vapor deposition and, for our experiment, pressurized in a diamond anvil cell. Under hydrostatic pressure we observe blue-shifts of the SiV and GeV zero-phonon lines by 17 THz (70 meV) and 78 THz (320 meV), respectively. These measured pressure induced shifts are in good agreement with ab initio calculations that take into account the lattice compression based on the equation of state of diamond and that are extended to the case of the tin-vacancy (SnV) center. This work provides guidance on the use of group-IV-vacancy centers as quantum sensors under extreme pressures that will exploit their specific optical and spin properties induced by their intrinsic inversion-symmetric structure., Comment: 7 pages, 4 figures

We propose world value functions (WVFs), a type of goal-oriented general value function that represents how to solve not just a given task, but any other goal-reaching task in an agent's environment. This is achieved by equipping an agent with an internal goal space defined as all the world states where it experiences a terminal transition. The agent can then modify the standard task rewards to define its own reward function, which provably drives it to learn how to achieve all reachable internal goals, and the value of doing so in the current task. We demonstrate two key benefits of WVFs in the context of learning and planning. In particular, given a learned WVF, an agent can compute the optimal policy in a new task by simply estimating the task's reward function. Furthermore, we show that WVFs also implicitly encode the transition dynamics of the environment, and so can be used to perform planning. Experimental results show that WVFs can be learned faster than regular value functions, while their ability to infer the environment's dynamics can be used to integrate learning and planning methods to further improve sample efficiency., Comment: Accepted at the Planning and Reinforcement Learning Workshop at ICAPS 2022. arXiv admin note: text overlap with arXiv:2205.08827

In this paper, we prove that the self-dual morphological hierarchical structure computed on a n-D gray-level wellcomposed image u by the algorithm of G{\'e}raud et al. [1] is exactly the mathematical structure defined to be the tree of shape of u in Najman et al [2]. We recall that this algorithm is in quasi-linear time and thus considered to be optimal. The tree of shapes leads to many applications in mathematical morphology and in image processing like grain filtering, shapings, image segmentation, and so on.

$L$-functions typically encode interesting information about mathematical objects. This paper reports 29 identities between such functions that hitherto never appeared in the literature. Of these we have a complete proof for 9; all others are extensively numerically checked and we welcome proofs of their (in)validity. The method we devised to obtain these identities is a two-step process whereby a list of candidate identities is automatically generated, obtained, tested, and ultimately formally proven. The approach is however only \emph{semi-}automated as human intervention is necessary for the post-processing phase, to determine the most general form of a conjectured identity and to provide a proof for them. This work complements other instances in the literature where automated symbolic computation has served as a productive step toward theorem proving and can be extended in several directions further to explore the algebraic landscape of $L$-functions and similar constructions.

In Mathematical Morphology (MM), connected filters based on dynamics are used to filter the extrema of an image. Similarly, persistence is a concept coming from Persistent Homology (PH) and Morse Theory (MT) that represents the stability of the extrema of a Morse function. Since these two concepts seem to be closely related, in this paper we examine their relationship, and we prove that they are equal on n-D Morse functions, n $\ge$ 1. More exactly, pairing a minimum with a 1-saddle by dynamics or pairing the same 1-saddle with a minimum by persistence leads exactly to the same pairing, assuming that the critical values of the studied Morse function are unique. This result is a step further to show how much topological data analysis and mathematical morphology are related, paving the way for a more in-depth study of the relations between these two research fields., Comment: Journal of Mathematical Imaging and Vision, Springer Verlag, In press

It is desirable for an agent to be able to solve a rich variety of problems that can be specified through language in the same environment. A popular approach towards obtaining such agents is to reuse skills learned in prior tasks to generalise compositionally to new ones. However, this is a challenging problem due to the curse of dimensionality induced by the combinatorially large number of ways high-level goals can be combined both logically and temporally in language. To address this problem, we propose a framework where an agent first learns a sufficient set of skill primitives to achieve all high-level goals in its environment. The agent can then flexibly compose them both logically and temporally to provably achieve temporal logic specifications in any regular language, such as regular fragments of linear temporal logic. This provides the agent with the ability to map from complex temporal logic task specifications to near-optimal behaviours zero-shot. We demonstrate this experimentally in a tabular setting, as well as in a high-dimensional video game and continuous control environment. Finally, we also demonstrate that the performance of skill machines can be improved with regular off-policy reinforcement learning algorithms when optimal behaviours are desired., Comment: Published as a conference paper at ICLR 2024

An open problem in artificial intelligence is how to learn and represent knowledge that is sufficient for a general agent that needs to solve multiple tasks in a given world. In this work we propose world value functions (WVFs), which are a type of general value function with mastery of the world - they represent not only how to solve a given task, but also how to solve any other goal-reaching task. To achieve this, we equip the agent with an internal goal space defined as all the world states where it experiences a terminal transition - a task outcome. The agent can then modify task rewards to define its own reward function, which provably drives it to learn how to achieve all achievable internal goals, and the value of doing so in the current task. We demonstrate a number of benefits of WVFs. When the agent's internal goal space is the entire state space, we demonstrate that the transition function can be inferred from the learned WVF, which allows the agent to plan using learned value functions. Additionally, we show that for tasks in the same world, a pretrained agent that has learned any WVF can then infer the policy and value function for any new task directly from its rewards. Finally, an important property for long-lived agents is the ability to reuse existing knowledge to solve new tasks. Using WVFs as the knowledge representation for learned tasks, we show that an agent is able to solve their logical combination zero-shot, resulting in a combinatorially increasing number of skills throughout their lifetime., Comment: Accepted to the 5th Multi-disciplinary Conference on Reinforcement Learning and Decision Making (RLDM), 2022

We present a novel fiber-optic component, a "tapered multi-core fiber (MCF)", designed for integration into ultra-miniaturized endoscopes for minimally invasive two-photon point-scanning imaging and to address the power delivery issue that has faced MCF based lensless endoscopes. With it we achieve experimentally a factor 6.0 increase in two-photon signal yield while keeping the ability to point-scan by the memory effect, and a factor 8.9 sacrificing the memory effect. To reach this optimal design we first develop and validate a fast numerical model capable of predicting the essential properties of an arbitrarily tapered MCF from its structural parameters. We then use this model to identify the tapered MCF design parameters that result in a chosen set of target properties (point-spread function, delivered power, presence or absence of memory effect). We fabricate the identified target designs by stack-and-draw and post-processing on a CO$_{2}$ laser-based glass processing and splicing system. Finally we demonstrate the performance gain of the fabricated tapered MCFs in two-photon imaging when used in a lensless endoscope system. Our results show that tailoring of the taper profile brings new degrees of freedom that can be efficiently exploited for lensless endoscopes.

We report high efficiency, miniaturized, ultra-fast coherent beam combining with 3D printed micro-optics directly on the tip of a multicore fiber bundle. The highly compact device foot-print (180 micron diameter) facilitates its incorporation into a minimally invasive ultra-thin nonlinear endoscope to perform two-photon imaging, Comment: Published in Optics Letters

Two common approaches to sequential decision-making are AI planning (AIP) and reinforcement learning (RL). Each has strengths and weaknesses. AIP is interpretable, easy to integrate with symbolic knowledge, and often efficient, but requires an up-front logical domain specification and is sensitive to noise; RL only requires specification of rewards and is robust to noise but is sample inefficient and not easily supplied with external knowledge. We propose an integrative approach that combines high-level planning with RL, retaining interpretability, transfer, and efficiency, while allowing for robust learning of the lower-level plan actions. Our approach defines options in hierarchical reinforcement learning (HRL) from AIP operators by establishing a correspondence between the state transition model of AI planning problem and the abstract state transition system of a Markov Decision Process (MDP). Options are learned by adding intrinsic rewards to encourage consistency between the MDP and AIP transition models. We demonstrate the benefit of our integrated approach by comparing the performance of RL and HRL algorithms in both MiniGrid and N-rooms environments, showing the advantage of our method over the existing ones., Comment: 30 pages, 15 figures

Segmentation of curvilinear structures is important in many applications, such as retinal blood vessel segmentation for early detection of vessel diseases and pavement crack segmentation for road condition evaluation and maintenance. Currently, deep learning-based methods have achieved impressive performance on these tasks. Yet, most of them mainly focus on finding powerful deep architectures but ignore capturing the inherent curvilinear structure feature (e.g., the curvilinear structure is darker than the context) for a more robust representation. In consequence, the performance usually drops a lot on cross-datasets, which poses great challenges in practice. In this paper, we aim to improve the generalizability by introducing a novel local intensity order transformation (LIOT). Specifically, we transfer a gray-scale image into a contrast-invariant four-channel image based on the intensity order between each pixel and its nearby pixels along with the four (horizontal and vertical) directions. This results in a representation that preserves the inherent characteristic of the curvilinear structure while being robust to contrast changes. Cross-dataset evaluation on three retinal blood vessel segmentation datasets demonstrates that LIOT improves the generalizability of some state-of-the-art methods. Additionally, the cross-dataset evaluation between retinal blood vessel segmentation and pavement crack segmentation shows that LIOT is able to preserve the inherent characteristic of curvilinear structure with large appearance gaps. An implementation of the proposed method is available at https://github.com/TY-Shi/LIOT., Comment: Accepted by IEEE TIP

We propose a framework that learns to execute natural language instructions in an environment consisting of goal-reaching tasks that share components of their task descriptions. Our approach leverages the compositionality of both value functions and language, with the aim of reducing the sample complexity of learning novel tasks. First, we train a reinforcement learning agent to learn value functions that can be subsequently composed through a Boolean algebra to solve novel tasks. Second, we fine-tune a seq2seq model pretrained on web-scale corpora to map language to logical expressions that specify the required value function compositions. Evaluating our agent in the BabyAI domain, we observe a decrease of 86% in the number of training steps needed to learn a second task after mastering a single task. Results from ablation studies further indicate that it is the combination of compositional value functions and language representations that allows the agent to quickly generalize to new tasks., Comment: Presented at AI-HRI symposium as part of AAAI-FSS 2021 (arXiv:2109.10836)