Large language models (LLMs) can perform complex reasoning in few- and zero-shot settings by generating intermediate chain of thought (CoT) reasoning steps. Further, each reasoning step can rely on external tools to support computation beyond the core LLM capabilities (e.g. search/running code). Prior work on CoT prompting and tool use typically requires hand-crafting task-specific demonstrations and carefully scripted interleaving of model generations with tool use. We introduce Automatic Reasoning and Tool-use (ART), a framework that uses frozen LLMs to automatically generate intermediate reasoning steps as a program. Given a new task to solve, ART selects demonstrations of multi-step reasoning and tool use from a task library. At test time, ART seamlessly pauses generation whenever external tools are called, and integrates their output before resuming generation. ART achieves a substantial improvement over few-shot prompting and automatic CoT on unseen tasks in the BigBench and MMLU benchmarks, and matches performance of hand-crafted CoT prompts on a majority of these tasks. ART is also extensible, and makes it easy for humans to improve performance by correcting errors in task-specific programs or incorporating new tools, which we demonstrate by drastically improving performance on select tasks with minimal human intervention.

Context: Requirements engineering researchers have been experimenting with machine learning and deep learning approaches for a range of RE tasks, such as requirements classification, requirements tracing, ambiguity detection, and modelling. However, most of today's ML/DL approaches are based on supervised learning techniques, meaning that they need to be trained using a large amount of task-specific labelled training data. This constraint poses an enormous challenge to RE researchers, as the lack of labelled data makes it difficult for them to fully exploit the benefit of advanced ML/DL technologies. Objective: This paper addresses this problem by showing how a zero-shot learning approach can be used for requirements classification without using any labelled training data. We focus on the classification task because many RE tasks can be framed as classification problems. Method: The ZSL approach used in our study employs contextual word-embeddings and transformer-based language models. We demonstrate this approach through a series of experiments to perform three classification tasks: (1)FR/NFR: classification functional requirements vs non-functional requirements; (2)NFR: identification of NFR classes; (3)Security: classification of security vs non-security requirements. Results: The study shows that the ZSL approach achieves an F1 score of 0.66 for the FR/NFR task. For the NFR task, the approach yields F1~0.72-0.80, considering the most frequent classes. For the Security task, F1~0.66. All of the aforementioned F1 scores are achieved with zero-training efforts. Conclusion: This study demonstrates the potential of ZSL for requirements classification. An important implication is that it is possible to have very little or no training data to perform classification tasks. The proposed approach thus contributes to the solution of the long-standing problem of data shortage in RE.

Generative Adversarial Networks (GAN) have been widely investigated for image synthesis based on their powerful representation learning ability. In this work, we explore the StyleGAN and its application of synthetic food image generation. Despite the impressive performance of GAN for natural image generation, food images suffer from high intra-class diversity and inter-class similarity, resulting in overfitting and visual artifacts for synthetic images. Therefore, we aim to explore the capability and improve the performance of GAN methods for food image generation. Specifically, we first choose StyleGAN3 as the baseline method to generate synthetic food images and analyze the performance. Then, we identify two issues that can cause performance degradation on food images during the training phase: (1) inter-class feature entanglement during multi-food classes training and (2) loss of high-resolution detail during image downsampling. To address both issues, we propose to train one food category at a time to avoid feature entanglement and leverage image patches cropped from high-resolution datasets to retain fine details. We evaluate our method on the Food-101 dataset and show improved quality of generated synthetic food images compared with the baseline. Finally, we demonstrate the great potential of improving the performance of downstream tasks, such as food image classification by including high-quality synthetic training samples in the data augmentation.

In order to efficiently learn a dynamics model for a task in a new environment, one can adapt a model learned in a similar source environment. However, existing adaptation methods can fail when the target dataset contains transitions where the dynamics are very different from the source environment. For example, the source environment dynamics could be of a rope manipulated in free-space, whereas the target dynamics could involve collisions and deformation on obstacles. Our key insight is to improve data efficiency by focusing model adaptation on only the regions where the source and target dynamics are similar. In the rope example, adapting the free-space dynamics requires significantly fewer data than adapting the free-space dynamics while also learning collision dynamics. We propose a new method for adaptation that is effective in adapting to regions of similar dynamics. Additionally, we combine this adaptation method with prior work on planning with unreliable dynamics to make a method for data-efficient online adaptation, called FOCUS. We first demonstrate that the proposed adaptation method achieves statistically significantly lower prediction error in regions of similar dynamics on simulated rope manipulation and plant watering tasks. We then show on a bimanual rope manipulation task that FOCUS achieves data-efficient online learning, in simulation and in the real world.

In recent years, the training requirements of many state-of-the-art Deep Learning (DL) models have scaled beyond the compute and memory capabilities of a single processor, and necessitated distribution among processors. Training such massive models necessitates advanced parallelism strategies to maintain efficiency. However, such distributed DL parallelism strategies require a varied mixture of collective and point-to-point communication operations across a broad range of message sizes and scales. Examples of models using advanced parallelism strategies include Deep Learning Recommendation Models (DLRM) and Mixture-of-Experts (MoE). Communication libraries' performance varies wildly across different communication operations, scales, and message sizes. We propose MCR-DL: an extensible DL communication framework that supports all point-to-point and collective operations while enabling users to dynamically mix-and-match communication backends for a given operation without deadlocks. MCR-DL also comes packaged with a tuning suite for dynamically selecting the best communication backend for a given input tensor. We select DeepSpeed-MoE and DLRM as candidate DL models and demonstrate a 31% improvement in DS-MoE throughput on 256 V100 GPUs on the Lassen HPC system. Further, we achieve a 20% throughput improvement in a dense Megatron-DeepSpeed model and a 25% throughput improvement in DLRM on 32 A100 GPUs with the Theta-GPU HPC system.

Genito-Pelvic Pain/Penetration-Disorder (GPPPD) is a common disorder but rarely treated in routine care. Previous research documents that GPPPD symptoms can be treated effectively using internet-based psychological interventions. However, non-response remains common for all state-of-the-art treatments and it is unclear which patient groups are expected to benefit most from an internet-based intervention. Multivariable prediction models are increasingly used to identify predictors of heterogeneous treatment effects, and to allocate treatments with the greatest expected benefits. In this study, we developed and internally validated a multivariable decision tree model that predicts effects of an internet-based treatment on a multidimensional composite score of GPPPD symptoms. Data of a randomized controlled trial comparing the internet-based intervention to a waitlist control group (N =200) was used to develop a decision tree model using model-based recursive partitioning. Model performance was assessed by examining the apparent and bootstrap bias-corrected performance. The final pruned decision tree consisted of one splitting variable, joint dyadic coping, based on which two response clusters emerged. No effect was found for patients with low dyadic coping ($n$=33; $d$=0.12; 95% CI: -0.57-0.80), while large effects ($d$=1.00; 95%CI: 0.68-1.32; $n$=167) are predicted for those with high dyadic coping at baseline. The bootstrap-bias-corrected performance of the model was $R^2$=27.74% (RMSE=13.22).

OpenAI also announced new partnerships with language learning app Duolingo and Be My Eyes, an application for the visually impaired, to create AI Chatbots which can assist their users using natural language.

CAMBRIDGE, MA -- March 14, 2023 -- Researchers from MIT and other institutions report that a "hash function" -- a core database search operation -- can be significantly accelerated through the use of machine learning. The hope is that the new technique could accelerate computational systems that scientists use to store and analyze DNA, amino acid sequences, or other biological information. Hashing is used in applications from database indexing to data compression to cryptography. A hash function generates codes that directly determine the location where data would be stored. But because traditional hash functions generate codes randomly, sometimes two pieces of data can be hashed with the same value.

The question will be simple but perpetual: Person or machine? Every encounter with language, other than in the flesh, will now bring with it that small, consuming test. For some--teachers, professors, journalists--the question of humanity will be urgent and essential. For those who operate in the large bureaucratic apparatus of boilerplate--copywriters, lawyers, advertisers, political strategists--the question will be irrelevant except as a matter of efficiency. How will they use new artificial-intelligence technology to accelerate the production of language that was already mostly automatic? For everyone, the question will now hover, quotidian and cosmic, over words wherever you find them: Who's there?

To the CSE community: We are excited to introduce you to the newest member of our department -SmartWheels-, a miniature robotic vehicle! Our robotic vehicle is outfitted with the most advanced sensor technology, including LiDAR, radar, and cameras, enabling it to perceive its surroundings in real-time and make intelligent decisions based on its surroundings. It is also powered by cutting-edge AI algorithms that enable it to learn from its experiences and enhance its performance continuously. This platform will host numerous master's theses and graduation projects. We appreciate you taking the time to learn about our latest innovation.