Andela exists to connect brilliance and opportunity. Since 2014, we have been dedicated to breaking down global barriers and accelerating the future of work for both technologists and organizations around the world. For technologists, Andela offers competitive long term career opportunities with leading organizations, access to a global community of professionals, and education opportunities with leading technology providers. For companies, Andela provides access to a global network of fully integrated team members that unlock their business' innovation and growth potential. At Andela, we are deeply passionate about creating long-lasting and transformative growth opportunities for all and doing it in an E.P.I.C. [andela.com/careers]

Transformer models have recently emerged as one of the foundational models in natural language processing, and as a byproduct, there is significant recent interest and investment in scaling these models. However, the training and inference costs of these large Transformer language models are prohibitive, thus necessitating more research in identifying more efficient variants. In this work, we propose a simple yet effective modification to the Transformer architecture inspired by the literature in statistical language modeling, by augmenting the model with n-grams that are constructed from a discrete latent representation of the text sequence. We evaluate our model, the N-Grammer on language modeling on the C4 data-set as well as text classification on the SuperGLUE data-set, and find that it outperforms several strong baselines such as the Transformer and the Primer. We open-source our model for reproducibility purposes in Jax.

As demonstrated by GPT-3 and T5, transformers grow in capability as parameter spaces become larger and larger. However, for tasks that require a large amount of knowledge, non-parametric memory allows models to grow dramatically with a sub-linear increase in computational cost and GPU memory requirements. Recent models such as RAG and REALM have introduced retrieval into conditional generation. These models incorporate neural initial retrieval from a corpus of passages. We build on this line of research, proposing Re2G, which combines both neural initial retrieval and reranking into a BART-based sequence-to-sequence generation. Our reranking approach also permits merging retrieval results from sources with incomparable scores, enabling an ensemble of BM25 and neural initial retrieval. To train our system end-to-end, we introduce a novel variation of knowledge distillation to train the initial retrieval, reranker, and generation using only ground truth on the target sequence output. We find large gains in four diverse tasks: zero-shot slot filling, question answering, fact-checking, and dialog, with relative gains of 9% to 34% over the previous state-of-the-art on the KILT leaderboard. We make our code available as open source at https://github.com/IBM/kgi-slot-filling/tree/re2g.

Remote sensing is the process of detecting and monitoring the physical characteristics of an area by measuring its reflected and emitted radiation at a distance. It is being broadly used to monitor ecosystems, mainly for their preservation. Ever-growing reports of invasive species have affected the natural balance of ecosystems. Exotic invasive species have a critical impact when introduced into new ecosystems and may lead to the extinction of native species. In this study, we focus on Ludwigia peploides, considered by the European Union as an aquatic invasive species. Its presence can negatively impact the surrounding ecosystem and human activities such as agriculture, fishing, and navigation. Our goal was to develop a method to identify the presence of the species. We used images collected by a drone-mounted multispectral sensor to achieve this, creating our LudVision data set. To identify the targeted species on the collected images, we propose a new method for detecting Ludwigia p. in multispectral images. The method is based on existing state-of-the-art semantic segmentation methods modified to handle multispectral data. The proposed method achieved a producer's accuracy of 79.9% and a user's accuracy of 95.5%.

In order to promote the use of machine learning in 5G, the International Telecommunication Union (ITU) proposed in 2021 the second edition of the ITU AI/ML in 5G challenge, with over 1600 participants from 82 countries. This work details the second place solution overall, which is also the winning solution of the Graph Neural Networking Challenge 2021. We tackle the problem of generalization when applying a model to a 5G network that may have longer paths and larger link capacities than the ones observed in training. To achieve this, we propose to first extract robust features related to Queueing Theory (QT), and then fine-tune the analytical baseline prediction using a modification of the Routenet Graph Neural Network (GNN) model. The proposed solution generalizes much better than simply using Routenet, and manages to reduce the analytical baseline's 10.42 mean absolute percent error to 1.45 (1.27 with an ensemble). This suggests that making small changes to an approximate model that is known to be robust can be an effective way to improve accuracy without compromising generalization.

Textual knowledge bases such as Wikipedia require considerable effort to keep up to date and consistent. While automated writing assistants could potentially ease this burden, the problem of suggesting edits grounded in external knowledge has been under-explored. In this paper, we introduce the novel generation task of *faithfully reflecting updated information in text* (FRUIT) where the goal is to update an existing article given new evidence. We release the FRUIT-WIKI dataset, a collection of over 170K distantly supervised data produced from pairs of Wikipedia snapshots, along with our data generation pipeline and a gold evaluation set of 914 instances whose edits are guaranteed to be supported by the evidence. We provide benchmark results for popular generation systems as well as EDIT5 -- a T5-based approach tailored to editing we introduce that establishes the state of the art. Our analysis shows that developing models that can update articles faithfully requires new capabilities for neural generation models, and opens doors to many new applications.

A common sales strategy involves having account executives (AEs) actively reach out and contact potential customers. However, not all contact attempts have a positive effect: some attempts do not change customer decisions, while others might even interfere with the desired outcome. In this work we propose using causal inference to estimate the effect of contacting each potential customer and setting the contact policy accordingly. We demonstrate this approach on data from Worthy, an online jewelry marketplace. We examined the Worthy business process to identify relevant decisions and outcomes, and formalized assumptions on how they were made. Using causal tools, we selected a decision point where improving AE contact activity appeared to be promising. We then generated a personalized policy and recommended reaching out only to customers for whom it would be beneficial. Finally, we validated the results in an A\B test over a 3-month period, resulting in an increase in item delivery rate of the targeted population by 22% (p-value=0.026). This policy is now being used on an ongoing basis.

The availability of genomic data has grown exponentially in the last decade, mainly due to the development of new sequencing technologies. Based on the interactions between genes (and gene products) extracted from the increasing genomic data, numerous studies have focused on the identification of associations between genes and functions. While these studies have shown great promise, the problem of annotating genes with functions remains an open challenge. In this work, we present a method to detect missing annotations in hierarchical multi-label classification datasets. We propose a method that exploits the class hierarchy by computing aggregated probabilities to the paths of classes from the leaves to the root for each instance. The proposed method is presented in the context of predicting missing gene function annotations, where these aggregated probabilities are further used to select a set of annotations to be verified through in vivo experiments. The experiments on Oriza sativa Japonica, a variety of rice, showcase that incorporating the hierarchy of classes into the method often improves the predictive performance and our proposed method yields superior results when compared to competitor methods from the literature. Genomic data has become exponentially available in the last decade, mainly due to the development of new technologies, including gene expression profiling generated with RNA sequencing (Ranganathan, Gribskov, Nakai and Schönbach, 2019). Based on the interactions between genes (and gene products) extracted from the increasing genomic data, numerous studies have focused on the identification of associations between genes and functions (Rust, Mongin and Birney, 2002; Vandepoele, Quimbaya, Casneuf, De Veylder and Van de Peer, 2009; van Dam, Võsa, van der Graaf, Franke and de Magalhães, 2017).

Additionally, a general family of activation functions is proposed by considering both max-pooling and nonlinear operators in the context of morphological representations.

Analyzing the planet at scale with satellite imagery and machine learning is a dream that has been constantly hindered by the cost of difficult-to-access highly-representative high-resolution imagery. To remediate this, we introduce here the WorldStrat dataset. The largest and most varied such publicly available dataset, at Airbus SPOT 6/7 satellites' high resolution of up to 1.5 m/pixel, empowered by European Space Agency's Phi-Lab as part of the ESA-funded QueryPlanet project, we curate nearly 10,000 sqkm of unique locations to ensure stratified representation of all types of land-use across the world: from agriculture to ice caps, from forests to multiple urbanization densities. We also enrich those with locations typically under-represented in ML datasets: sites of humanitarian interest, illegal mining sites, and settlements of persons at risk. We temporally-match each high-resolution image with multiple low-resolution images from the freely accessible lower-resolution Sentinel-2 satellites at 10 m/pixel. We accompany this dataset with an open-source Python package to: rebuild or extend the WorldStrat dataset, train and infer baseline algorithms, and learn with abundant tutorials, all compatible with the popular EO-learn toolbox. We hereby hope to foster broad-spectrum applications of ML to satellite imagery, and possibly develop from free public low-resolution Sentinel2 imagery the same power of analysis allowed by costly private high-resolution imagery. We illustrate this specific point by training and releasing several highly compute-efficient baselines on the task of Multi-Frame Super-Resolution. High-resolution Airbus imagery is CC BY-NC, while the labels and Sentinel2 imagery are CC BY, and the source code and pre-trained models under BSD. The dataset is available at https://zenodo.org/record/6810792 and the software package at https://github.com/worldstrat/worldstrat .