Recently, prompt-tuning has attracted growing interests in event argument extraction (EAE). However, the existing prompt-tuning methods have not achieved satisfactory performance due to the lack of consideration of entity information. In this paper, we propose a bi-directional iterative prompt-tuning method for EAE, where the EAE task is treated as a cloze-style task to take full advantage of entity information and pre-trained language models (PLMs). Furthermore, our method explores event argument interactions by introducing the argument roles of contextual entities into prompt construction. Since template and verbalizer are two crucial components in a cloze-style prompt, we propose to utilize the role label semantic knowledge to construct a semantic verbalizer and design three kinds of templates for the EAE task. Experiments on the ACE 2005 English dataset with standard and low-resource settings show that the proposed method significantly outperforms the peer state-of-the-art methods. Our code is available at https://github.com/HustMinsLab/BIP.

Reliable neural networks (NNs) provide important inference-time reliability guarantees such as fairness and robustness. Complementarily, privacy-preserving NN inference protects the privacy of client data. So far these two emerging areas have been largely disconnected, yet their combination will be increasingly important. In this work, we present the first system which enables privacy-preserving inference on reliable NNs. Our key idea is to design efficient fully homomorphic encryption (FHE) counterparts for the core algorithmic building blocks of randomized smoothing, a state-of-the-art technique for obtaining reliable models. The lack of required control flow in FHE makes this a demanding task, as na\"ive solutions lead to unacceptable runtime. We employ these building blocks to enable privacy-preserving NN inference with robustness and fairness guarantees in a system called Phoenix. Experimentally, we demonstrate that Phoenix achieves its goals without incurring prohibitive latencies. To our knowledge, this is the first work which bridges the areas of client data privacy and reliability guarantees for NNs.

Context: Machine learning software can generate models that inappropriately discriminate against specific protected social groups (e.g., groups based on gender, ethnicity, etc). Motivated by those results, software engineering researchers have proposed many methods for mitigating those discriminatory effects. While those methods are effective in mitigating bias, few of them can provide explanations on what is the root cause of bias. Objective: We aim at better detection and mitigation of algorithmic discrimination in machine learning software problems. Method: Here we propose xFAIR, a model-based extrapolation method, that is capable of both mitigating bias and explaining the cause. In our xFAIR approach, protected attributes are represented by models learned from the other independent variables (and these models offer extrapolations over the space between existing examples). We then use the extrapolation models to relabel protected attributes later seen in testing data or deployment time. Our approach aims to offset the biased predictions of the classification model via rebalancing the distribution of protected attributes. Results: The experiments of this paper show that, without compromising (original) model performance, xFAIR can achieve significantly better group and individual fairness (as measured in different metrics) than benchmark methods. Moreover, when compared to another instance-based rebalancing method, our model-based approach shows faster runtime and thus better scalability. Conclusion: Algorithmic decision bias can be removed via extrapolation that smooths away outlier points. As evidence for this, our proposed xFAIR is not only performance-wise better (measured by fairness and performance metrics) than two state-of-the-art fairness algorithms.

The ability to identify influential training examples enables us to debug training data and explain model behavior. Existing techniques to do so are based on the flow of training data influence through the model parameters. For large models in NLP applications, it is often computationally infeasible to study this flow through all model parameters, therefore techniques usually pick the last layer of weights. However, we observe that since the activation connected to the last layer of weights contains "shared logic", the data influenced calculated via the last layer weights prone to a ``cancellation effect'', where the data influence of different examples have large magnitude that contradicts each other. The cancellation effect lowers the discriminative power of the influence score, and deleting influential examples according to this measure often does not change the model's behavior by much. To mitigate this, we propose a technique called TracIn-WE that modifies a method called TracIn to operate on the word embedding layer instead of the last layer, where the cancellation effect is less severe. One potential concern is that influence based on the word embedding layer may not encode sufficient high level information. However, we find that gradients (unlike embeddings) do not suffer from this, possibly because they chain through higher layers. We show that TracIn-WE significantly outperforms other data influence methods applied on the last layer significantly on the case deletion evaluation on three language classification tasks for different models. In addition, TracIn-WE can produce scores not just at the level of the overall training input, but also at the level of words within the training input, a further aid in debugging.

The 4th Industrial Revolution is the culmination of the digital age. Nowadays, technologies such as robotics, nanotechnology, genetics, and artificial intelligence promise to transform our world and the way we live. Artificial Intelligence Ethics and Safety is an emerging research field that has been gaining popularity in recent years. Several private, public and non-governmental organizations have published guidelines proposing ethical principles for regulating the use and development of autonomous intelligent systems. Meta-analyses of the AI Ethics research field point to convergence on certain principles that supposedly govern the AI industry. However, little is known about the effectiveness of this form of Ethics. In this paper, we would like to conduct a critical analysis of the current state of AI Ethics and suggest that this form of governance based on principled ethical guidelines is not sufficient to norm the AI industry and its developers. We believe that drastic changes are necessary, both in the training processes of professionals in the fields related to the development of software and intelligent systems and in the increased regulation of these professionals and their industry. To this end, we suggest that law should benefit from recent contributions from bioethics, to make the contributions of AI ethics to governance explicit in legal terms.

The European Parliament's co-rapporteurs of the AI Act have proposed expanding the European Commission's revision powers to extend the list of high-risk systems and prohibited practices at a later stage. The eighth batch of compromise amendments on the proposed Artificial Intelligence regulation was shared by leading lawmakers Dragoș Tudorache and Brando Benifei with the representatives of the other political groups last Friday (21 October). The AI Act is a flagship legislative proposal to regulate the development, deployment and use of artificial intelligence. The co-rapporteurs are currently trying to reach a common position with the other political groups, pitching several significant changes to the original wording. The most relevant amendment would significantly extend the Commission's revision powers after the legislation becomes effective.

Oct 25 - Tesla Inc (TSLA.O) is under criminal investigation in the United States over claims that the company's electric vehicles can drive themselves, three people familiar with the matter said. The U.S. Department of Justice launched the previously undisclosed probe last year following more than a dozen crashes, some of them fatal, involving Tesla's driver assistance system Autopilot, which was activated during the accidents, the people said. As early as 2016, Tesla's marketing materials have touted Autopilot's capabilities. On a conference call that year, Elon Musk, the Silicon Valley automaker's chief executive, described it as "probably better" than a human driver. Last week, Musk said on another call Tesla would soon release an upgraded version of "Full Self-Driving" software allowing customers to travel "to your work, your friend's house, to the grocery store without you touching the wheel."

Tesla is under criminal investigation in the United States over claims that the company's electric vehicles can drive themselves, three people familiar with the matter said. The US Department of Justice (DoJ) launched the previously undisclosed investigation last year following more than a dozen crashes, some of them fatal, involving Tesla's driver assistance system known as Autopilot, which was activated during the accidents, the people said. As early as 2016, Tesla's marketing materials have touted Autopilot's capabilities. On a conference call that year, Elon Musk, the Tesla chief executive, described it as "probably better" than a human driver. Last week, Musk said on another call Tesla would soon release an upgraded version of "full self-driving" software, allowing customers to travel "to your work, your friend's house, to the grocery store without you touching the wheel".

Want to have an impressionist painting of Thai temples in the style of Claude Monet, but you cannot afford to commission an artist? Let artificial intelligence (AI) do the work for you. Then you change your mind and want to have the painting in a surrealistic style. Type what you want in the message field of the AI art-generating program. You get what you wanted.

Dr. Stephen Thaler developed DABUS (Device for Autonomous Bootstrapping of Unified Science), an artificial intelligence (AI) system that can autonomously create patentable inventions. Thaler has filed patent applications in various jurisdictions for two inventions created by DABUS – a food container with side walls having a fractal profile, and a beacon for attracting enhanced attention for example in a search and rescue scenario[1]. In each application, Thaler listed DABUS as the sole inventor, forcing patent offices in various jurisdictions to address the issue of whether an AI system can be an inventor on a patent application. Thus far, the DABUS patent applications have found very limited success in patent offices and courts around the world. In the latest decision, the United States Court of Appeals for the Federal Circuit (CAFC) held that the US Patent Act requires an inventor to be a natural person, and consequently, an AI system cannot be an inventor on a United States patent application.[2] The DABUS applications were initially rejected by the United States Patent and Trademark Office (USPTO).