A central but unresolved aspect of problem-solving in AI is the capability to introduce and use abstractions, something humans excel at. Work in cognitive science has demonstrated that humans tend towards higher levels of abstraction when engaged in collaborative task-oriented communication, enabling gradually shorter and more information-efficient utterances. Several computational methods have attempted to replicate this phenomenon, but all make unrealistic simplifying assumptions about how abstractions are introduced and learned. Our method, Abstractions for Communicating Efficiently (ACE), overcomes these limitations through a neuro-symbolic approach. On the symbolic side, we draw on work from library learning for proposing abstractions. We combine this with neural methods for communication and reinforcement learning, via a novel use of bandit algorithms for controlling the exploration and exploitation trade-off in introducing new abstractions. ACE exhibits similar tendencies to humans on a collaborative construction task from the cognitive science literature, where one agent (the architect) instructs the other (the builder) to reconstruct a scene of block-buildings. ACE results in the emergence of an efficient language as a by-product of collaborative communication. Beyond providing mechanistic insights into human communication, our work serves as a first step to providing conversational agents with the ability for human-like communicative abstractions.

This article is part of a VB special issue. Enterprise companies have experimented with artificial intelligence (AI) for years -- a pilot here, a use case there. But company leaders have long dreamed of going bigger, better and faster when it comes to AI. That is, applying AI at scale. The goals of this quest may vary.

The results of a new study by Domino Data Lab confirm the importance of data science and advanced analytics to modern enterprises. Upside: A recent survey conducted by Domino Data Lab found that nearly four in five respondents agreed that data science, ML, and AI are critical to the overall future growth of their company. In fact, 36 percent said these were the single most critical factors. Kjell Carlsson: It's worth noting that these numbers are similar to the surveys I've run in the past. I did a survey at Forrester in 2021, where 25 percent said data science was the single most important factor for their competitiveness and expected that to rise to 51 percent in the next two years.

Modern enterprises are inundated with data; however, not all data is usable as is for machine learning. Though an organization may have millions of data points, it could still have data struggles that stunt machine learning. Turning to synthetic data for machine learning can boost privacy, democratize data, minimize bias in data sets and reduce costs. More broadly, real data and synthetic data tend to be used in combination. "I can't think of any project in the AI space where you wouldn't be able to get a better outcome by leveraging synthetic data," said Kjell Carlsson, principal analyst at Forrester Research.

Topological deep learning is a formalism that is aimed at introducing topological language to deep learning for the purpose of utilizing the minimal mathematical structures to formalize problems that arise in a generic deep learning problem. This is the first of a sequence of articles with the purpose of introducing and studying this formalism. In this article, we define and study the classification problem in machine learning in a topological setting. Using this topological framework, we show when the classification problem is possible or not possible in the context of neural networks. Finally, we show that for a given data, the architecture of a classification neural network must take into account the topology of this data in order to achieve a successful classification task.

Learning disentangled representations is regarded as a fundamental task for improving the generalization, robustness, and interpretability of generative models. However, measuring disentanglement has been challenging and inconsistent, often dependent on an ad-hoc external model or specific to a certain dataset. To address this, we present a method for quantifying disentanglement that only uses the generative model, by measuring the topological similarity of conditional submanifolds in the learned representation. To illustrate the effectiveness and applicability of our method, we empirically evaluate several state-of-the-art models across multiple datasets. We find that our method ranks models similarly to existing methods. Figure 1: Factors in the dSprites dataset displaying topological similarity and semantic correspondence to respective latent dimensions in a disentangled generative model, as shown through Wasserstein RLT distributions of homology and latent interpolations along respective dimensions. Learning disentangled representations is important for a variety of tasks, including adversarial robustness, generalization to novel tasks, and interpretability (Stutz et al., 2019; Alemi et al., 2017; Ridgeway, 2016; Bengio et al., 2013). Recently, deep generative models have shown marked improvement in disentanglement across an increasing number of datasets and a variety of training objectives (Chen et al., 2016; Lin et al., 2020; Higgins et al., 2017; Kim and Mnih, 2018; Chen et al., 2018b; Burgess et al., 2018; Karras et al., 2019). Nevertheless, quantifying the extent of this disentanglement has remained challenging and inconsistent.

If your organization is implementing or thinking of implementing a contact-tracing app, it's wise to consider more than just workforce safety. Failing to do so could expose your company other risks such as employment-related lawsuits and compliance issues. More fundamentally, companies should be thinking about the ethical implications of their AI use. Contact-tracing apps are raising a lot of questions. For example, should employers be able to use them? If so, must employees opt-in or can employers make them mandatory?

We introduce giotto-tda, a Python library that integrates high-performance topological data analysis with machine learning via a scikit-learn-compatible API and state-of-the-art C++ implementations. The library's ability to handle various types of data is rooted in a wide range of preprocessing techniques, and its strong focus on data exploration and interpretability is aided by an intuitive plotting API. Source code, binaries, examples, and documentation can be found at https://github.com/giotto-ai/giotto-tda.

Forrester Research recently released a report entitled, Shatter the Seven Myths of Machine Learning. In it, the authors warn, "Unfortunately, there is a pandemic of ML misconceptions and literacy among business leaders who must make critical decisions about ML projects." When executives and managers talk about AI and machine learning, they sometimes make factual mistakes that reveal their true level of knowledge. Forrester senior analyst Kjell Carlsson, who is the lead author of the report, said in a recent interview that he's heard audible sighs over the phone when experts hear what lay people have to say. "When the head of product says something like, 'We're using reinforcement learning because we're incorporating user feedback into the trends modeling,' that's probably not a good thing," said Carlsson.

Understanding how neural networks learn remains one of the central challenges in machine learning research. From random at the start of training, the weights of a neural network evolve in such a way as to be able to perform a variety of tasks, like classifying images. Here we study the emergence of structure in the weights by applying methods from topological data analysis. We train simple feedforward neural networks on the MNIST dataset and monitor the evolution of the weights. When initialized to zero, the weights follow trajectories that branch off recurrently, thus generating trees that describe the growth of the effective capacity of each layer. When initialized to tiny random values, the weights evolve smoothly along two-dimensional surfaces. We show that natural coordinates on these learning surfaces correspond to important factors of variation.