Verbal communication plays a crucial role in human cooperation, particularly when the partners only have incomplete information about the task, environment, and each other's mental state. In this paper, we propose a novel cooperative communication framework, Goal-Oriented Mental Alignment (GOMA). GOMA formulates verbal communication as a planning problem that minimizes the misalignment between the parts of agents' mental states that are relevant to the goals. This approach enables an embodied assistant to reason about when and how to proactively initialize communication with humans verbally using natural language to help achieve better cooperation. We evaluate our approach against strong baselines in two challenging environments, Overcooked (a multiplayer game) and VirtualHome (a household simulator). Our experimental results demonstrate that large language models struggle with generating meaningful communication that is grounded in the social and physical context. In contrast, our approach can successfully generate concise verbal communication for the embodied assistant to effectively boost the performance of the cooperation as well as human users' perception of the assistant.

LLMs can generate factually incorrect statements even when provided access to reference documents. Such errors can be dangerous in high-stakes applications (e.g., document-grounded QA for healthcare or finance). We present GenAudit -- a tool intended to assist fact-checking LLM responses for document-grounded tasks. GenAudit suggests edits to the LLM response by revising or removing claims that are not supported by the reference document, and also presents evidence from the reference for facts that do appear to have support. We train models to execute these tasks, and design an interactive interface to present suggested edits and evidence to users. Comprehensive evaluation by human raters shows that GenAudit can detect errors in 8 different LLM outputs when summarizing documents from diverse domains. To ensure that most errors are flagged by the system, we propose a method that can increase the error recall while minimizing impact on precision. We release our tool (GenAudit) and fact-checking model for public use.

In this study, we introduce BEnQA, a dataset comprising parallel Bengali and English exam questions for middle and high school levels in Bangladesh. Our dataset consists of approximately 5K questions covering several subjects in science with different types of questions, including factual, application, and reasoning-based questions. We benchmark several Large Language Models (LLMs) with our parallel dataset and observe a notable performance disparity between the models in Bengali and English. We also investigate some prompting methods, and find that Chain-of-Thought prompting is beneficial mostly on reasoning questions, but not so much on factual ones. We also find that appending English translation helps to answer questions in Bengali. Our findings point to promising future research directions for improving the performance of LLMs in Bengali and more generally in low-resource languages.

Zero-shot generalization (ZSG) to unseen dynamics is a major challenge for creating generally capable embodied agents. To address the broader challenge, we start with the simpler setting of contextual reinforcement learning (cRL), assuming observability of the context values that parameterize the variation in the system's dynamics, such as the mass or dimensions of a robot, without making further simplifying assumptions about the observability of the Markovian state. Toward the goal of ZSG to unseen variation in context, we propose the contextual recurrent state-space model (cRSSM), which introduces changes to the world model of the Dreamer (v3) (Hafner et al., 2023). This allows the world model to incorporate context for inferring latent Markovian states from the observations and modeling the latent dynamics. Our experiments show that such systematic incorporation of the context improves the ZSG of the policies trained on the ``dreams'' of the world model. We further find qualitatively that our approach allows Dreamer to disentangle the latent state from context, allowing it to extrapolate its dreams to the many worlds of unseen contexts. The code for all our experiments is available at \url{https://github.com/sai-prasanna/dreaming_of_many_worlds}.

Standard language models generate text by selecting tokens from a fixed, finite, and standalone vocabulary. We introduce a novel method that selects context-aware phrases from a collection of supporting documents. One of the most significant challenges for this paradigm shift is determining the training oracles, because a string of text can be segmented in various ways and each segment can be retrieved from numerous possible documents. To address this, we propose to initialize the training oracles using linguistic heuristics and, more importantly, bootstrap the oracles through iterative self-reinforcement. Extensive experiments show that our model not only outperforms standard language models on a variety of knowledge-intensive tasks but also demonstrates improved generation quality in open-ended text generation. For instance, compared to the standard language model counterpart, our model raises the accuracy from 23.47% to 36.27% on OpenbookQA, and improves the MAUVE score from 42.61% to 81.58% in open-ended text generation. Remarkably, our model also achieves the best performance and the lowest latency among several retrieval-augmented baselines. In conclusion, we assert that retrieval is more accurate generation and hope that our work will encourage further research on this new paradigm shift.

Legal research is a crucial task in the practice of law. It requires intense human effort and intellectual prudence to research a legal case and prepare arguments. Recent boom in generative AI has not translated to proportionate rise in impactful legal applications, because of low trustworthiness and and the scarcity of specialized datasets for training Large Language Models (LLMs). This position paper explores the potential of LLMs within Legal Text Analytics (LTA), highlighting specific areas where the integration of human expertise can significantly enhance their performance to match that of experts. We introduce a novel dataset and describe a human centered, compound AI system that principally incorporates human inputs for performing LTA tasks with LLMs.

Recent development of large language models (LLMs) has exhibited impressive zero-shot proficiency on generic and common sense questions. However, LLMs' application on domain-specific vertical questions still lags behind, primarily due to the humiliation problems and deficiencies in vertical knowledge. Furthermore, the vertical data annotation process often requires labor-intensive expert involvement, thereby presenting an additional challenge in enhancing the model's vertical capabilities. In this paper, we propose SERVAL, a synergy learning pipeline designed for unsupervised development of vertical capabilities in both LLMs and small models by mutual enhancement. Specifically, SERVAL utilizes the LLM's zero-shot outputs as annotations, leveraging its confidence to teach a robust vertical model from scratch. Reversely, the trained vertical model guides the LLM fine-tuning to enhance its zero-shot capability, progressively improving both models through an iterative process. In medical domain, known for complex vertical knowledge and costly annotations, comprehensive experiments show that, without access to any gold labels, SERVAL with the synergy learning of OpenAI GPT-3.5 and a simple model attains fully-supervised competitive performance across ten widely used medical datasets. These datasets represent vertically specialized medical diagnostic scenarios (e.g., diabetes, heart diseases, COVID-19), highlighting the potential of SERVAL in refining the vertical capabilities of LLMs and training vertical models from scratch, all achieved without the need for annotations.

Artificial General Intelligence falls short when communicating role specific nuances to other systems. This is more pronounced when building autonomous LLM agents capable and designed to communicate with each other for real world problem solving. Humans can communicate context and domain specific nuances along with knowledge, and that has led to refinement of skills. In this work we propose and evaluate a novel method that leads to knowledge distillation among LLM agents leading to realtime human role play preserving unique contexts without relying on any stored data or pretraining. We also evaluate how our system performs better in simulated real world tasks compared to state of the art.

A new automated humanoid robot powered by OpenAI's ChatGPT resembles something akin to the AI Skynet from the sci-fi film Terminator While the new robot is not a killing machine, Figure 01 can perform basic autonomous tasks and carry out real-time conversations with humans - with the help of ChatGPT. The company, Figure AI, shared a demonstration video, showing how ChatGPT helps the two-legged machine visual objects, plan future actions and even reflect on its memory. Figure's cameras snap its surrounding and send them to a a large vision-language model trained by OpenAI, which than translates the images back to the robot. The clip showed a man asking the humanoid to put away dirty laundry, wash dishes and hand him something to eat - and the robot performed the tasks - but unlike ChatGPT, Figure is more hesitant when it comes to answering questions. Figure AI hopes that its first AI humanoid robot will prove capable at jobs too dangerous for human laborers and might alleviate worker shortages. 'Two weeks ago, we announced Figure OpenAI are joining forces to push the boundaries of robot learning,' Figure founder Brett Adcock wrote on X. 'Together we are developing next-generation AI models for our humanoid robots,' he added.

Around 2014, researchers working on AI realized that GPUs worked better for training AI than other chips. While the more-common computer processing units, or CPUs, are optimized to handle a small number of highly complicated tasks, GPUs are better at doing the math for many simple calculations at the same time. Because training AI requires making many connections between billions of different words or images, GPUs worked well. Many of the major AI breakthroughs that helped set the foundation for the large language models at the center of today's AI revolution were made by running AI algorithms on Nvidia GPUs.