It's no secret that artificial intelligence, algorithms, and big data have a problem with gender and racial bias. These systems can be biased based on who builds them, how they're developed, and how they're ultimately used. Trying to solve the problem is a community of Black data scientists, researchers, and organizations. This article highlights the Black women amongst their ranks, who are exposing algorithmic biases, empowering communities of color with data, and arguing for more diverse representation. Joy Buolamwini is a Ghanaian-American computer scientist based at MIT Media Lab.

The following week, she took part in several workshops at NeurIPS, the largest annual AI research conference, which over 20,000 people attended this year. It was "therapeutic," she says, to see how the community she'd helped build showed up and supported one another. Now, another week later, she's just winding down and catching her breath--and trying to make sense of it all. On Monday, December 14, I caught up with Gebru via Zoom. She recounted what happened during her time at Google, reflected on what it meant for the field and AI ethics research, and gave parting words of advice to those who want to keep holding tech companies accountable.

Artificial intelligence has made great strides since the deep learning revolution, but AI systems still struggle to extrapolate outside of their training data and adapt to new situations. For inspiration we look to the domain of science, where scientists have been able to develop theories which show remarkable ability to extrapolate and sometimes predict the existence of phenomena which have never been observed before. According to David Deutsch, this type of extrapolation, which he calls "reach", is due to scientific theories being hard to vary. In this work we investigate Deutsch's hard-to-vary principle and how it relates to more formalized principles in deep learning such as the bias-variance trade-off and Occam's razor. We distinguish internal variability, how much a model/theory can be varied internally while still yielding the same predictions, with external variability, which is how much a model must be varied to accurately predict new, out-of-distribution data. We discuss how to measure internal variability using the size of the Rashomon set and how to measure external variability using Kolmogorov complexity. We explore what role hard-to-vary explanations play in intelligence by looking at the human brain and distinguish two learning systems in the brain. The first system operates similar to deep learning and likely underlies most of perception and motor control while the second is a more creative system capable of generating hard-to-vary explanations of the world. We argue that figuring out how replicate this second system, which is capable of generating hard-to-vary explanations, is a key challenge which needs to be solved in order to realize artificial general intelligence. We make contact with the framework of Popperian epistemology which rejects induction and asserts that knowledge generation is an evolutionary process which proceeds through conjecture and refutation.

Such a query can be answered currently at a high human effort cost, by inspecting e.g., a JSON list of Assemblée elected officials (available from NosDeputes.fr) and manually connecting the names with those found in a national registry of companies. This considerable effort may still miss connections that could be found if one added information about politicians' and business people's spouses, information sometimes available in public knowledge bases such as DBPedia, or journalists' notes. No single query language can be used on such heterogeneous data; instead, we study methods to query the corpus by specifying some keywords and asking for all the connections that exist, in one or across several data sources, between these keywords. This problem has been studied under the name of keyword search over structured data, in particular for relational databases [49, 27], XML documents [24, 33], RDF graphs [30, 16]. However, most of these works assumed one single source of data, in which connections among nodes are clearly identified. When authors considered several data sources [31], they still assumed that one query answer comes from a single data source. In contrast, the ConnectionLens system [10] answers keyword search queries over arbitrary combinations of datasets and heterogeneous data models, independently produced by actors unaware of each other's existence.

Analog hardware implemented deep learning models are promising for computation and energy constrained systems such as edge computing devices. However, the analog nature of the device and the associated many noise sources will cause changes to the value of the weights in the trained deep learning models deployed on such devices. In this study, systematic evaluation of the inference performance of trained popular deep learning models for image classification deployed on analog devices has been carried out, where additive white Gaussian noise has been added to the weights of the trained models during inference. It is observed that deeper models and models with more redundancy in design such as VGG are more robust to the noise in general. However, the performance is also affected by the design philosophy of the model, the detailed structure of the model, the exact machine learning task, as well as the datasets.

Given a time-evolving tensor with missing entries, how can we effectively factorize it for precisely predicting the missing entries? Tensor factorization has been extensively utilized for analyzing various multi-dimensional real-world data. However, existing models for tensor factorization have disregarded the temporal property for tensor factorization while most real-world data are closely related to time. Moreover, they do not address accuracy degradation due to the sparsity of time slices. The essential problems of how to exploit the temporal property for tensor decomposition and consider the sparsity of time slices remain unresolved. In this paper, we propose TATD (Time-Aware Tensor Decomposition), a novel tensor decomposition method for real-world temporal tensors. TATD is designed to exploit temporal dependency and time-varying sparsity of real-world temporal tensors. We propose a new smoothing regularization with Gaussian kernel for modeling time dependency. Moreover, we improve the performance of TATD by considering time-varying sparsity. We design an alternating optimization scheme suitable for temporal tensor factorization with our smoothing regularization. Extensive experiments show that TATD provides the state-of-the-art accuracy for decomposing temporal tensors.

One of the key problems in tensor completion is the number of uniformly random sample entries required for recovery guarantee. The main aim of this paper is to study $n_1 \times n_2 \times n_3$ third-order tensor completion and investigate into incoherence conditions of $n_3$ low-rank $n_1$-by-$n_2$ matrix slices under the transformed tensor singular value decomposition where the unitary transformation is applied along $n_3$-dimension. We show that such low-rank tensors can be recovered exactly with high probability when the number of randomly observed entries is of order $O( r\max \{n_1, n_2 \} \log ( \max \{ n_1, n_2 \} n_3))$, where $r$ is the sum of the ranks of these $n_3$ matrix slices in the transformed tensor. By utilizing synthetic data and imaging data sets, we demonstrate that the theoretical result can be obtained under valid incoherence conditions, and the tensor completion performance of the proposed method is also better than that of existing methods in terms of sample sizes requirement.

Consider a prediction setting where a few inputs (e.g., satellite images) are expensively annotated with the prediction targets (e.g., crop types), and many inputs are cheaply annotated with auxiliary information (e.g., climate information). How should we best leverage this auxiliary information for the prediction task? Empirically across three image and time-series datasets, and theoretically in a multi-task linear regression setting, we show that (i) using auxiliary information as input features improves in-distribution error but can hurt out-of-distribution (OOD) error; while (ii) using auxiliary information as outputs of auxiliary tasks to pre-train a model improves OOD error. To get the best of both worlds, we introduce In-N-Out, which first trains a model with auxiliary inputs and uses it to pseudolabel all the in-distribution inputs, then pre-trains a model on OOD auxiliary outputs and fine-tunes this model with the pseudolabels (self-training). We show both theoretically and empirically that In-N-Out outperforms auxiliary inputs or outputs alone on both in-distribution and OOD error.

Earlier this fall, A.I. ethicist Timnit Gebru submitted a paper for consideration at an academic conference about predictive language models: on their environmental cost, and how they could learn racist and sexist language and also spread misinformation. Since she was working for Google, the company first wanted to review the paper--which Gebru wrote with several of her colleagues--and sign off on it. She was then told by senior managers that the paper didn't meet Google's publication bar, and that she should retract it or remove the names of Google employees. Gebru wanted more clarity on why they wanted it retracted and said that if Google couldn't provide that information, she would resign. This kicked off a few days of wrangling and several intense emails--until a manager emailed Gebru's boss, saying they had accepted her resignation.

The fourth-annual Midas List Europe, produced by Forbes in partnership with TrueBridge Capital Partners, has arrived, and we're excited to share the top companies that drove the portfolios of this year's top European venture capitalists. The outlook for the European venture market may have been cloudy at the beginning of the global pandemic as recessionary cutbacks loomed and the IPO window narrowed, but European startups and investors have since bounced back. A wide variety of tech-based startups have been able to ride the tailwinds of the crisis, with new areas of everyday life benefitting from the transition to a technology-driven environment. Evidentially, investors remain clear-eyed and eager to invest in growth and innovation on either side of the pond with European VC deal value – and potentially fundraising – on pace to set new annual records. Here are the top ten companies that acted as key drivers behind this year's Midas List Europe: It's been a boom year for Stockholm-based Spotify, which is making its third consecutive appearance as the #1 driver on the Midas List Europe and fourth appearance overall.