The Forward Forward algorithm, proposed by Geoffrey Hinton in November 2022, is a novel method for training neural networks as an alternative to backpropagation. In this project, we replicate Hinton's experiments on the MNIST dataset, and subsequently extend the scope of the method with two significant contributions. First, we establish a baseline performance for the Forward Forward network on the IMDb movie reviews dataset. As far as we know, our results on this sentiment analysis task marks the first instance of the algorithm's extension beyond computer vision. Second, we introduce a novel pyramidal optimization strategy for the loss threshold - a hyperparameter specific to the Forward Forward method. Our pyramidal approach shows that a good thresholding strategy causes a difference of up to 8% in test error. Lastly, we perform visualizations of the trained parameters and derived several significant insights, such as a notably larger (10-20x) mean and variance in the weights acquired by the Forward Forward network. Repository: https://github.com/Ads-cmu/ForwardForward

Understanding how the statistical and geometric properties of neural activity relate to performance is a key problem in theoretical neuroscience and deep learning. Here, we calculate how correlations between object representations affect the capacity, a measure of linear separability. We show that for spherical object manifolds, introducing correlations between centroids effectively pushes the spheres closer together, while introducing correlations between the axes effectively shrinks their radii, revealing a duality between correlations and geometry with respect to the problem of classification. We then apply our results to accurately estimate the capacity of deep network data.

Ethics in AI has become a debated topic of public and expert discourse in recent years. But what do people who build AI - AI practitioners - have to say about their understanding of AI ethics and the challenges associated with incorporating it in the AI-based systems they develop? Understanding AI practitioners' views on AI ethics is important as they are the ones closest to the AI systems and can bring about changes and improvements. We conducted a survey aimed at understanding AI practitioners' awareness of AI ethics and their challenges in incorporating ethics. Based on 100 AI practitioners' responses, our findings indicate that majority of AI practitioners had a reasonable familiarity with the concept of AI ethics, primarily due to workplace rules and policies. Privacy protection and security was the ethical principle that majority of them were aware of. Formal education/training was considered somewhat helpful in preparing practitioners to incorporate AI ethics. The challenges that AI practitioners faced in the development of ethical AI-based systems included (i) general challenges, (ii) technology-related challenges and (iii) human-related challenges. We also identified areas needing further investigation and provided recommendations to assist AI practitioners and companies in incorporating ethics into AI development.

During the past decade, deep learning has transformed a number of historically challenging problems in computer vision, natural language processing, game intelligence, etc. In many of these applications, the trained neural networks used to solve these problems are over-parameterized. That is, the neural networks have far more parameters than the number of data points used for training. In this setting, a neural network can typically fit any training data - including random labels [95] - making it hard to explain why deep learning methods generalize so well [36]. Moreover, the practical performance of neural networks often improves as the number of parameters grow [55,84]. These observations have led to the study of the potential implicit regularization (sometimes called implicit bias) imposed by the gradient based methods and different network architectures [8, 68, 69]. It may seem surprising that there is a link to generalized hardness of approximation (GHA), as this phenomenon - at a first glance - may seem disconnected from implicit regularization. However, the GHA phenomenon (see 1.2), which first appeared in [13] (see also [2] Chapter 8) and analyzed [13, 34, 41] in connection with robust and convex optimization [20, 21, 63, 64], typically stem from regularization problems (e.g.

With the emergence of multimodal electronic health records, the evidence for an outcome may be captured across multiple modalities ranging from clinical to imaging and genomic data. Predicting outcomes effectively requires fusion frameworks capable of modeling fine-grained and multi-faceted complex interactions between modality features within and across patients. We develop an innovative fusion approach called MaxCorr MGNN that models non-linear modality correlations within and across patients through Hirschfeld-Gebelein-Renyi maximal correlation (MaxCorr) embeddings, resulting in a multi-layered graph that preserves the identities of the modalities and patients. We then design, for the first time, a generalized multi-layered graph neural network (MGNN) for task-informed reasoning in multi-layered graphs, that learns the parameters defining patient-modality graph connectivity and message passing in an end-to-end fashion. We evaluate our model an outcome prediction task on a Tuberculosis (TB) dataset consistently outperforming several state-of-the-art neural, graph-based and traditional fusion techniques.

To foster the development of new models for collaborative AI-assisted report generation, we introduce MegaWika, consisting of 13 million Wikipedia articles in 50 diverse languages, along with their 71 million referenced source materials. We process this dataset for a myriad of applications, going beyond the initial Wikipedia citation extraction and web scraping of content, including translating non-English articles for cross-lingual applications and providing FrameNet parses for automated semantic analysis. MegaWika is the largest resource for sentence-level report generation and the only report generation dataset that is multilingual. We manually analyze the quality of this resource through a semantically stratified sample. Finally, we provide baseline results and trained models for crucial steps in automated report generation: cross-lingual question answering and citation retrieval.

The dominant text generation models compose the output by sequentially selecting words from a fixed vocabulary. In this paper, we formulate text generation as progressively copying text segments (e.g., words or phrases) from an existing text collection. We compute the contextualized representations of meaningful text segments and index them using efficient vector search toolkits. The task of text generation is then decomposed into a series of copy-and-paste operations: at each time step, we seek suitable text spans from the text collection rather than selecting from a standalone vocabulary. Experiments on the standard language modeling benchmark (WikiText-103) show that our approach achieves better generation quality according to both automatic and human evaluations. Besides, its inference efficiency is comparable to token-level autoregressive models thanks to the reduction of decoding steps. We also show that our approach allows for effective domain adaptation by simply switching to domain-specific text collection without extra training. Finally, we observe that our approach attains additional performance gains by simply scaling up to larger text collections, again without further training. Specifically, LMs generate the next-token distribution over a fixed vocabulary for any given prefix.

In this study, we address the challenge of using energy-based models to produce high-quality, label-specific data in complex structured datasets, such as population genetics, RNA or protein sequences data. Traditional training methods encounter difficulties due to inefficient Markov chain Monte Carlo mixing, which affects the diversity of synthetic data and increases generation times. To address these issues, we use a novel training algorithm that exploits non-equilibrium effects. This approach, applied on the Restricted Boltzmann Machine, improves the model's ability to correctly classify samples and generate high-quality synthetic data in only a few sampling steps. The effectiveness of this method is demonstrated by its successful application to four different types of data: handwritten digits, mutations of human genomes classified by continental origin, functionally characterized sequences of an enzyme protein family, and homologous RNA sequences from specific taxonomies.

Ageing detection and failure prediction are essential in many Internet of Things (IoT) deployments, which operate huge quantities of embedded devices unattended in the field for years. In this paper, we present a large-scale empirical analysis of natural SRAM wear-out using 154 boards from a general-purpose testbed. Starting from SRAM initialization bias, which each node can easily collect at startup, we apply various metrics for feature extraction and experiment with common machine learning methods to predict the age of operation for this node. Our findings indicate that even though ageing impacts are subtle, our indicators can well estimate usage times with an $R^2$ score of 0.77 and a mean error of 24% using regressors, and with an F1 score above 0.6 for classifiers applying a six-months resolution.

Accurate and efficient classification of different types of cancer is critical for early detection and effective treatment. In this paper, we present the results of our experiments using the EfficientNet algorithm for classification of brain tumor, breast cancer mammography, chest cancer, and skin cancer. We used publicly available datasets and preprocessed the images to ensure consistency and comparability. Our experiments show that the EfficientNet algorithm achieved high accuracy, precision, recall, and F1 scores on each of the cancer datasets, outperforming other state-of-the-art algorithms in the literature. We also discuss the strengths and weaknesses of the EfficientNet algorithm and its potential applications in clinical practice. Our results suggest that the EfficientNet algorithm is well-suited for classification of different types of cancer and can be used to improve the accuracy and efficiency of cancer diagnosis.