A key challenge in social network analysis is understanding the position, or stance, of people in the graph on a large set of topics. While past work has modeled (dis)agreement in social networks using signed graphs, these approaches have not modeled agreement patterns across a range of correlated topics. For instance, disagreement on one topic may make disagreement(or agreement) more likely for related topics. We propose the Stance Embeddings Model(SEM), which jointly learns embeddings for each user and topic in signed social graphs with distinct edge types for each topic. By jointly learning user and topic embeddings, SEM is able to perform cold-start topic stance detection, predicting the stance of a user on topics for which we have not observed their engagement. We demonstrate the effectiveness of SEM using two large-scale Twitter signed graph datasets we open-source. One dataset, TwitterSG, labels (dis)agreements using engagements between users via tweets to derive topic-informed, signed edges. The other, BirdwatchSG, leverages community reports on misinformation and misleading content. On TwitterSG and BirdwatchSG, SEM shows a 39% and 26% error reduction respectively against strong baselines.

While a lot of research in explainable AI focuses on producing effective explanations, less work is devoted to the question of how people understand and interpret the explanation. In this work, we focus on this question through a study of saliency-based explanations over textual data. Feature-attribution explanations of text models aim to communicate which parts of the input text were more influential than others towards the model decision. Many current explanation methods, such as gradient-based or Shapley value-based methods, provide measures of importance which are well-understood mathematically. But how does a person receiving the explanation (the explainee) comprehend it? And does their understanding match what the explanation attempted to communicate? We empirically investigate the effect of various factors of the input, the feature-attribution explanation, and visualization procedure, on laypeople's interpretation of the explanation. We query crowdworkers for their interpretation on tasks in English and German, and fit a GAMM model to their responses considering the factors of interest. We find that people often mis-interpret the explanations: superficial and unrelated factors, such as word length, influence the explainees' importance assignment despite the explanation communicating importance directly. We then show that some of this distortion can be attenuated: we propose a method to adjust saliencies based on model estimates of over- and under-perception, and explore bar charts as an alternative to heatmap saliency visualization. We find that both approaches can attenuate the distorting effect of specific factors, leading to better-calibrated understanding of the explanation.

Reasoning over natural language is a long-standing goal for the research community. However, studies have shown that existing language models are inadequate in reasoning. To address the issue, we present POET, a new pre-training paradigm. Through pre-training language models with programs and their execution results, POET empowers language models to harvest the reasoning knowledge possessed in program executors via a data-driven approach. POET is conceptually simple and can be instantiated by different kinds of programs. In this paper, we show three empirically powerful instances, i.e., POET-Math, POET-Logic, and POET-SQL. Experimental results on six benchmarks demonstrate that POET can significantly boost model performance on natural language reasoning, such as numerical reasoning, logical reasoning, and multi-hop reasoning. Taking the DROP benchmark as a representative example, POET improves the F1 metric of BART from 69.2% to 80.6%. Furthermore, POET shines in giant language models, pushing the F1 metric of T5-11B to 87.6% and achieving a new state-of-the-art performance on DROP. POET opens a new gate on reasoning-enhancement pre-training and we hope our analysis would shed light on the future research of reasoning like program executors.

As the sun rises over the island of American Samoa, a chorus of animal voices drifts upward. They're not the calls of birds, though -- the purrs, clicks and groans are coming from under the water. New research shows how automation can make it increasingly easy to eavesdrop on the fish making the sounds and uncover how their environment impacts them. Jill Munger first heard about fish that make sounds while she was an undergraduate student. A veteran researcher told her about marine acoustics.

Nvidia's GeForce RTX 3050 delivers great 1080p gaming performance with modern features, including capable ray tracing chops and DLSS. It has plenty of memory and doesn't make any unusual technical compromises, unlike AMD's rival Radeon RX 6500 XT, but that potentially makes it a target for GPU miners--which could mean bad things for price and availability. A year and a half into the latest generation of graphics cards--one plagued by chip shortages, logistics woes, tariffs, crypto demand, and scalpers--we're finally starting to see the first GPUs for PC gamers on a tighter budget. And as the GeForce RTX 3050 we're reviewing today shows, Nvidia and AMD couldn't be going about it any more differently. AMD landed the first strike. The Radeon RX 6500 XT arrived just last week, and AMD made some hard compromises to hit its low $199 price point.

Imagine a hurricane has just devastated your town – and flattened your home. Before you could even begin to consider rebuilding, you'd wait weeks – or even months – for a property assessor just to visit to take a look at the damage, let alone unlock the funds you need to get back on your feet. But what if you could take pictures of the damage on the day of the hurricane, using your smartphone, and upload them – and when your insurer receives them, they deposit the funds into your bank account on the same day? If this sounds like the future, it's not – it's happening now. Real-world impact In September 2021, we at Tractable launched our estimating technology for the home, in Japan.

I look back at my accomplishments in 2021. I worked on various applied ML projects from climate change, protecting sharks to art & design. Highlights include open-source projects, TensorFlow, computer vision, deep learning, art and design. As always, I have been busy learning and sharing my knowledge with the community! Early 2022 I worked as an ML researcher on a Frontier Development Lab project using ML for climate change (MLCC).

Microsoft continues to make significant investments in deep learning, computer vision, and AI. The Microsoft Maps Team has been leveraging that investment to identify map features at scale and produce high-quality building footprint data sets with the overall goal to add to the OpenStreetMap and MissingMaps humanitarian efforts. As of this post, the following locations are available and Microsoft offers access to this data under the Open Data Commons Open Database License (ODbL). Country/Region Million buildings United States of America 129.6 Nigeria and Kenya 50.5 South America 44.5 Uganda and Tanzania 17.9 Canada 11.8 Australia 11.3 As you might expect, the vintage of the footprints depends on the collection date of the underlying imagery. Bing Maps Imagery is a composite of multiple sources with different capture dates (ranging 2012 to 2021).

Pair-wise loss is an approach to metric learning that learns a semantic embedding by optimizing a loss function that encourages images from the same semantic class to be mapped closer than images from different classes. The literature reports a large and growing set of variations of the pair-wise loss strategies. Here we decompose the gradient of these loss functions into components that relate to how they push the relative feature positions of the anchor-positive and anchor-negative pairs. This decomposition allows the unification of a large collection of current pair-wise loss functions. Additionally, explicitly constructing pair-wise gradient updates to separate out these effects gives insights into which have the biggest impact, and leads to a simple algorithm that beats the state of the art for image retrieval on the CAR, CUB and Stanford Online products datasets.

When explaining AI behavior to humans, how is the communicated information being comprehended by the human explainee, and does it match what the explanation attempted to communicate? When can we say that an explanation is explaining something? We aim to provide an answer by leveraging theory of mind literature about the folk concepts that humans use to understand behavior. We establish a framework of social attribution by the human explainee, which describes the function of explanations: the concrete information that humans comprehend from them. Specifically, effective explanations should be coherent (communicate information which generalizes to other contrast cases), complete (communicating an explicit contrast case, objective causes, and subjective causes), and interactive (surfacing and resolving contradictions to the generalization property through iterations). We demonstrate that many XAI mechanisms can be mapped to folk concepts of behavior. This allows us to uncover their modes of failure that prevent current methods from explaining effectively, and what is necessary to enable coherent explanations.