Figure 1: (a) Parameter trajectories from (white) RL, (black) constrained RL, and (brown) ROGER on a simulated reward landscape with their (transparent) explorations. Brighter regions indicate higher rewards, while darker regions indicate lower rewards. The red areas highlight violations with a red dashed line indicating the constraint threshold. RL and constrained RL consistently violate constraints, possibly during exploration, while ROGER effectively avoids violations. A video of this experiment is available at https://youtu.be/Cqu7vL T Piw?si=jtzJCpRubbFHx06w. Abstract--Existing robot locomotion learning techniques rely heavily on the offline selection of proper reward weighting gains and cannot guarantee constraint satisfaction (i.e., constraint violation) during training. Thus, this work aims to address both issues by proposing Reward-Oriented Gains via Embodied Regulation (ROGER), which adapts reward-weighting gains online based on penalties received throughout the embodied interaction process. The ratio between the positive reward (primary reward) and negative reward (penalty) gains is automatically reduced as the learning approaches the constraint thresholds to avoid violation. Conversely, the ratio is increased when learning is in safe states to prioritize performance. With a 60-kg quadruped robot, ROGER achieved near-zero constraint violation throughout multiple learning trials. It also achieved up to 50% more primary reward than the equivalent state-of-the-art techniques. In MuJoCo continuous locomotion benchmarks, including a single-leg hopper, ROGER exhibited comparable or up to 100% higher performance and 60% less torque usage and orientation deviation compared to those trained with the default reward function.

Chain-of-thought (CoT) reasoning boosts large language models' (LLMs) performance on complex tasks but faces two key limitations: a lack of reliability when solely relying on LLM-generated reasoning chains and lower reasoning performance from natural language prompts compared with code prompts. To address these issues, we propose CoT-RAG, a novel reasoning framework with three key designs: (i) Knowledge Graph-driven CoT Generation, featuring knowledge graphs to modulate reasoning chain generation of LLMs, thereby enhancing reasoning credibility; (ii) Learnable Knowledge Case-aware RAG, which incorporates retrieval-augmented generation (RAG) into knowledge graphs to retrieve relevant sub-cases and sub-descriptions, providing LLMs with learnable information; (iii) Pseudo Program Prompting Execution, which promotes greater logical rigor by guiding LLMs to execute reasoning tasks as pseudo-programs. Evaluations on nine public datasets spanning three reasoning tasks reveal significant accuracy gains-ranging from 4.0% to 44.3%-over state-of-the-art methods. Furthermore, tests on four domain-specific datasets demonstrate exceptional accuracy and efficient execution, underscoring its practical applicability and scalability. Our code and data are available at https: //github.com/hustlfy123/CoT-RAG.

In our previous research, we provided a reasoning system (called LeSAC) based on argumentation theory to provide legal support to designers during the design process. Building on this, this paper explores how to provide designers with effective explanations for their legally relevant design decisions. We extend the previous system for providing explanations by specifying norms and the key legal or ethical principles for justifying actions in normative contexts. Considering that first-order logic has strong expressive power, in the current paper we adopt a first-order deontic logic system with deontic operators and preferences. We illustrate the advantages and necessity of introducing deontic logic and designing explanations under LeSAC by modelling two cases in the context of autonomous driving. In particular, this paper also discusses the requirements of the updated LeSAC to guarantee rationality, and proves that a well-defined LeSAC can satisfy the rationality postulate for rule-based argumentation frameworks. This ensures the system's ability to provide coherent, legally valid explanations for complex design decisions.

Daniela Rus, director of MIT’s Computer Science and Artificial Intelligence Lab (CSAIL), isn’t worried that robots will take over the world. Instead, she envisions robots and humans teaming up and achieving things that neither could do alone. In this excerpt from her new book, The Heart and the Chip: Our Bright Future with Robots, she explores how they can extend the reach of human capabilities.

Blockers can take a tiny task and stretch it over days or weeks. Taking a moment at the beginning of a project to look for and prevent possible blockers can improve productivity. These examples of personal, team, and organizational levels show how gathering the right information and performing preflight checks can save hours of wasted time later. Two IT workers--Andrew and Bertie (not their real names)--were assigned the same task. While this task should have taken about an hour of hands-on keyboard work, it took Andrew four days. Andrew began the task one sunny Monday morning. Work went well until he hit a speed bump and needed to ask the requester (who we will call Roger) a question. Andrew tried to find him on the company chat system, only to learn Roger was out of the office. Andrew sent an email instead.

Maria Schrader’s film, starring Dan Stevens as a robot designed to be the perfect man, confirms comedy as the playground of philosophy: nothing is funnier or more stirring than the sight of somebody learning how to be.

In this paper, we present a new objective prediction model for synthetic speech naturalness. It can be used to evaluate Text-To-Speech or Voice Conversion systems and works language independently. The model is trained end-to-end and based on a CNN-LSTM network that previously showed to give good results for speech quality estimation. We trained and tested the model on 16 different datasets, such as from the Blizzard Challenge and the Voice Conversion Challenge. Further, we show that the reliability of deep learning-based naturalness prediction can be improved by transfer learning from speech quality prediction models that are trained on objective POLQA scores. The proposed model is made publicly available and can, for example, be used to evaluate different TTS system configurations.

A variety of financial services companies have started to incorporate artificial intelligence (AI) into their operations-- ranging from quantitative asset managers that use machine learning (ML) models to predict price movements in securities to roboadvisor systems that use AI to help investors decide on their asset allocation. More broadly, companies are increasingly using AI to both analyze structured data, (e.g., asset flows, performance) and extract information from unstructured/alternative data (e.g., images, documents, social media posts) through image recognition and natural language understanding capabilities. The greater volume of data, along with AI and ML tools that can provide automated insights and analytics, offers significant opportunities for asset owners and asset managers to increase operational productivity, improve cybersecurity and manage risk, among other benefits. Currently, more than half of asset managers are in the early stages of AI initiatives, according to a Sapient Global Markets survey.1 And almost one-quarter of asset owners who invest in hedge funds use alternative data and big data analytics/AI to support their investment processes, according to an EY/Greenwich Associates survey.2

A YOUNG MAN, let's call him Roger, arrives at the emergency department complaining of belly pain and nausea. A physical exam reveals that the pain is focused in the lower right portion of his abdomen. The doctor worries that it could be appendicitis. But by the time the imaging results come back, Roger is feeling better, and the scan shows that his appendix appears normal. The doctor turns to the computer to prescribe two medications, one for nausea and Tylenol for pain, before discharging him. This is one of the fictitious scenarios presented to 55 physicians around the country as part of a study to look at the usability of electronic health records (EHRs).

Although Einstein himself made many advances in physics, from special and... [ ] general relativity to the photoelectric effect and statistical mechanics, there were many problems he couldn't solve during his life. How much better could AI have done? At the dawn of the 20th century, there were a number of crises in physics. Radiating objects like stars emitted a finite, well-defined amount of energy at every wavelength, defying the best predictions of the day. Newton's laws of motion broke down and failed when objects approached the speed of light.