Software powered by artificial intelligence (AI) already carries out multiple legal tasks today. In the future, AI will affect and own even more of these tasks. While consumer facing AI is rapidly evolving, AI in legal is still in its early days. Autonomous legal departments and law firms are still far away. Today, AI can conduct research, review documents, expedite digital discovery, aid due diligence and more, but it cannot replace advocacy or negotiation or appear in court.

In a rehearsal room at London's Young Vic theater last week, three dramatists were arguing with an artificial intelligence about how to write a play. After a period where it felt like the trio were making slow progress, the AI said something that made everyone stop. "If you want a computer to write a play, go and buy one. It won't need any empathy, it won't need any understanding," it said. "The computer will write a play that is for itself. It will be a play that will bore you to death."

Apple has released yet more details on its new photo-scanning features, as the controversy over whether they should be added to the iPhone continues. Earlier this month, Apple announced that it would be adding three new features to iOS, all of which are intended to fight against child sexual exploitation and the distribution of abuse imagery. One adds new information to Siri and search, another checks messages sent to children to see if they might contain inappropriate images, and the third compares photos on an iPhone with a database of known child sexual abuse material (CSAM) and alerts Apple if it is found. It is the latter of those three features that has proven especially controversial. Critics say that the feature is in contravention of Apple's commitment to privacy, and that it could in the future be used to scan for other kinds of images, such as political pictures on the phones of people living in authoritarian regimes.

The past decade has seen a substantial rise in the amount of mis- and disinformation online, from targeted disinformation campaigns to influence politics, to the unintentional spreading of misinformation about public health. This development has spurred research in the area of automatic fact checking, from approaches to detect check-worthy claims and determining the stance of tweets towards claims, to methods to determine the veracity of claims given evidence documents. These automatic methods are often content-based, using natural language processing methods, which in turn utilise deep neural networks to learn higher-order features from text in order to make predictions. As deep neural networks are black-box models, their inner workings cannot be easily explained. At the same time, it is desirable to explain how they arrive at certain decisions, especially if they are to be used for decision making. While this has been known for some time, the issues this raises have been exacerbated by models increasing in size, and by EU legislation requiring models to be used for decision making to provide explanations, and, very recently, by legislation requiring online platforms operating in the EU to provide transparent reporting on their services. Despite this, current solutions for explainability are still lacking in the area of fact checking. This thesis presents my research on automatic fact checking, including claim check-worthiness detection, stance detection and veracity prediction. Its contributions go beyond fact checking, with the thesis proposing more general machine learning solutions for natural language processing in the area of learning with limited labelled data. Finally, the thesis presents some first solutions for explainable fact checking.

Spiking neural networks (SNNs) have shown a potential for having low energy with unsupervised learning capabilities due to their biologically-inspired computation. However, they may suffer from accuracy degradation if their processing is performed under the presence of hardware-induced faults in memories, which can come from manufacturing defects or voltage-induced approximation errors. Since recent works still focus on the fault-modeling and random fault injection in SNNs, the impact of memory faults in SNN hardware architectures on accuracy and the respective fault-mitigation techniques are not thoroughly explored. Toward this, we propose ReSpawn, a novel framework for mitigating the negative impacts of faults in both the off-chip and on-chip memories for resilient and energy-efficient SNNs. The key mechanisms of ReSpawn are: (1) analyzing the fault tolerance of SNNs; and (2) improving the SNN fault tolerance through (a) fault-aware mapping (FAM) in memories, and (b) fault-aware training-and-mapping (FATM). If the training dataset is not fully available, FAM is employed through efficient bit-shuffling techniques that place the significant bits on the non-faulty memory cells and the insignificant bits on the faulty ones, while minimizing the memory access energy. Meanwhile, if the training dataset is fully available, FATM is employed by considering the faulty memory cells in the data mapping and training processes. The experimental results show that, compared to the baseline SNN without fault-mitigation techniques, ReSpawn with a fault-aware mapping scheme improves the accuracy by up to 70% for a network with 900 neurons without retraining.

This conversation with the composer Philip Glass and me discusses an exciting project in partnership with OpenAi, in which we trained a neural net on a corpus of Glass' work. He offers commentary on the music created by "his AI", as well as insights on composition and creating art. We then talk about the different limitations and capacities of humans and Artificial Intelligence–if and how neural nets can help us create art, appreciate art, and find the same things humans find meaningful. Due to the covid-19 pandemic, this call took place over video conference in December 2020. Art and tech are both captivating to me because they frame the elevation and the limitations of being human. Art is also closely intertwined with technological advancements, as movement shifting art seems predicated on tech. For example, the photography of Martin Munkacsi from the 1920s and 1930s revolutionized the art, as he is often credited for being the first photographer to explore dynamic and candid styles. The emergence and ability of these new forms of creation coincided with the technological advancements at the time that enabled flash and faster shutters–candid and spontaneous movement shots wouldn't have been technically possible to make with the cameras that existed before. The advancements in machine learning today, likewise, excite me for the possibilities and new forms in art and creation. The goal of this project is to explore the capacities of artificial intelligence as a new medium (or instrument or tool?) for art, and to create a collaborative music composition with Philip Glass and "his AI." More details about the project can be found below. Philip: Nice to see you.

For many learning problems one may not have access to fine grained label information; e.g., an image can be labeled as husky, dog, or even animal depending on the expertise of the annotator. In this work, we formalize these settings and study the problem of learning from such coarse data. Instead of observing the actual labels from a set $\mathcal{Z}$, we observe coarse labels corresponding to a partition of $\mathcal{Z}$ (or a mixture of partitions). Our main algorithmic result is that essentially any problem learnable from fine grained labels can also be learned efficiently when the coarse data are sufficiently informative. We obtain our result through a generic reduction for answering Statistical Queries (SQ) over fine grained labels given only coarse labels. The number of coarse labels required depends polynomially on the information distortion due to coarsening and the number of fine labels $|\mathcal{Z}|$. We also investigate the case of (infinitely many) real valued labels focusing on a central problem in censored and truncated statistics: Gaussian mean estimation from coarse data. We provide an efficient algorithm when the sets in the partition are convex and establish that the problem is NP-hard even for very simple non-convex sets.

Social media such as Instagram and Twitter have become important platforms for marketing and selling illicit drugs. Detection of online illicit drug trafficking has become critical to combat the online trade of illicit drugs. However, the legal status often varies spatially and temporally; even for the same drug, federal and state legislation can have different regulations about its legality. Meanwhile, more drug trafficking events are disguised as a novel form of advertising commenting leading to information heterogeneity. Accordingly, accurate detection of illicit drug trafficking events (IDTEs) from social media has become even more challenging. In this work, we conduct the first systematic study on fine-grained detection of IDTEs on Instagram. We propose to take a deep multimodal multilabel learning (DMML) approach to detect IDTEs and demonstrate its effectiveness on a newly constructed dataset called multimodal IDTE(MM-IDTE). Specifically, our model takes text and image data as the input and combines multimodal information to predict multiple labels of illicit drugs. Inspired by the success of BERT, we have developed a self-supervised multimodal bidirectional transformer by jointly fine-tuning pretrained text and image encoders. We have constructed a large-scale dataset MM-IDTE with manually annotated multiple drug labels to support fine-grained detection of illicit drugs. Extensive experimental results on the MM-IDTE dataset show that the proposed DMML methodology can accurately detect IDTEs even in the presence of special characters and style changes attempting to evade detection.

We rely on experts all the time. If you need financial advice, you ask an expert. If you are sick, you visit a doctor, and as a juror you may listen to an expert witness. In the future, however, artificial intelligence (AI) might replace many of these people. In forensic science, the expert witness plays a vital role.

Artificial Intelligence is Human Intelligence. Artificial Intelligence is a misnomer. ‘Artificial’ does correctly imply that it is an artifact. Something created by something else, more specifically by humans.