Despite the high demand for manually annotated image data, managing complex and costly annotation projects remains under-discussed. This is partly due to the fact that leading such projects requires dealing with a set of diverse and interconnected challenges which often fall outside the expertise of specific domain experts, leaving practical guidelines scarce. These challenges range widely from data collection to resource allocation and recruitment, from mitigation of biases to effective training of the annotators. This paper provides a domain-agnostic preparation guide for annotation projects, with a focus on scientific imagery. Drawing from the authors' extensive experience in managing a large manual annotation project, it addresses fundamental concepts including success measures, annotation subjects, project goals, data availability, and essential team roles. Additionally, it discusses various human biases and recommends tools and technologies to improve annotation quality and efficiency. The goal is to encourage further research and frameworks for creating a comprehensive knowledge base to reduce the costs of manual annotation projects across various fields.

Physics-informed neural networks approach the approximation of differential equations by directly incorporating their structure and given conditions in a loss function. This enables conditions like, e.g., invariants to be easily added during the modelling phase. In addition, the approach can be considered as mesh free and can be utilised to compute solutions on arbitrary grids after the training phase. Therefore, physics-informed neural networks are emerging as a promising alternative to solving differential equations with methods from numerical mathematics. However, their performance highly depends on a large variety of factors. In this paper, we systematically investigate and evaluate a core component of the approach, namely the training point distribution. We test two ordinary and two partial differential equations with five strategies for training data generation and shallow network architectures, with one and two hidden layers. In addition to common distributions, we introduce sine-based training points, which are motivated by the construction of Chebyshev nodes. The results are challenged by using certain parameter combinations like, e.g., random and fixed-seed weight initialisation for reproducibility. The results show the impact of the training point distributions on the solution accuracy and we find evidence that they are connected to the characteristics of the differential equation.

Distributed machine learning training and inference is common today because today's large models require more memory and compute than can be provided by a single GPU. Distributed models are generally produced by programmers who take a sequential model specification and apply several distribution strategies to distribute state and computation across GPUs. Unfortunately, bugs can be introduced in the process, and a distributed model implementation's outputs might differ from the sequential model's outputs. In this paper, we describe an approach to statically identify such bugs by checking model refinement, that is, can the sequential model's outputs be reconstructed from the distributed model's outputs? Our approach, implemented in GraphGuard, uses iterative rewriting to prove model refinement. Our approach can scale to today's large models and deployments: we evaluate it using GPT and Llama-3. Further, it provides actionable output that aids in bug localization.

In this paper, we present a reliable, scalable, time deterministic, model-free procedure to tune swarms of Micro Aerial Vehicles (MAVs) using basic sensory data. Two approaches to taking advantage of parallel tuning are presented. First, the tuning with averaging of the results on the basis of performance indices reported from the swarm with identical gains to decrease the negative effect of the noise in the measurements. Second, the tuning with parallel testing of varying set of gains across the swarm to reduce the tuning time. The presented methods were evaluated both in simulation and real-world experiments. The achieved results show the ability of the proposed approach to improve the results of the tuning while decreasing the tuning time, ensuring at the same time a reliable tuning mechanism.

In this brief paper, we present a naive aggregation algorithm for a typical learning problem with expert advice setting, in which the task of improving generalization, i.e., model validation, is embedded in the learning process as a sequential decision-making problem. In particular, we consider a class of learning problem of point estimations for modeling high-dimensional nonlinear functions, where a group of experts update their parameter estimates using the discrete-time version of gradient systems, with small additive noise term, guided by the corresponding subsample datasets obtained from the original dataset. Here, our main objective is to provide conditions under which such an algorithm will sequentially determine a set of mixing distribution strategies used for aggregating the experts' estimates that ultimately leading to an optimal parameter estimate, i.e., as a consensus solution for all experts, which is better than any individual expert's estimate in terms of improved generalization or learning performances. Finally, as part of this work, we present some numerical results for a typical case of nonlinear regression problem.

To remind you with a visual, here, in figure 1, is the whole pipeline. I've circled the part of the system we'll address in this article. You may ask, what exactly is a pipeline and why do we use one, whether for ML production or any programmatic production operation which is managed by orchestration? You typically use pipelines when the job, such as training or other operation handled by orchestration, has multiple steps that occur in sequential order: do step A, do step B, and like this. Putting the steps into a ML production pipeline has multiple benefits.

Distributed training in TensorFlow is built around data parallelism, where we can replicate the same model architecture on multiple devices and run different slices of input data on them. Here the device is nothing but a unit of CPU + GPU or separate units of GPUs and TPUs. This method follows like; our entire data is divided into equal numbers of slices. These slices are decided based on available devices to train; following each slice, there is a model to train on that slice.

TPUs (Tensor Processing Units) are application-specific integrated circuits (ASICs) that are optimized specifically for processing matrices. Google Colab provides experimental support for TPUs for free! In this article, we'll be discussing how to train a model using TPU on Colab. Specifically, we'll be training BERT for text classification using the transformers package by huggingface on a TPU. Since the TPU is optimized for some specific operations, we need to check if our model actually uses them; i.e. we need to check if the TPU actually helps our model to train faster.

It's been nearly 4 years since Tensorflow was released, and the library has evolved to its official second version. Tensorflow is Google's library for deep learning and artificial intelligence. Deep Learning has been responsible for some amazing achievements recently, such as: Generating beautiful, photo-realistic images of people and things that never existed (GANs) Beating world champions in the strategy game Go, and complex video games like CS:GO and Dota 2 (Deep Reinforcement Learning) Self-driving cars (Computer Vision) Speech recognition (e.g. Siri) and machine translation (Natural Language Processing) Even creating videos of people doing and saying things they never did (DeepFakes - a potentially nefarious application of deep learning) Tensorflow is the world's most popular library for deep learning, and it's built by Google, whose parent Alphabet recently became the most cash-rich company in the world (just a few days before I wrote this). It is the library of choice for many companies doing AI and machine learning.

Did you know that Colab includes the ability to select a free Cloud TPU for training models? That's right, a whole TPU for you to use all by yourself in a notebook! As of TensorFlow 1.11, you can train Keras models with TPUs. In this post, let's take a look at what changes you need to make to your code to be able to train a Keras model on TPUs. Note that some of this may be simplified even further with the release of TensorFlow 2.0 later this year, but I thought it'd be helpful to share these tips in case you'd like to try this out now.