There is an eerie emptiness at the seat of US economic power. The US Treasury is in shutdown like much of the federal government. Most staff are furloughed as the world's finance ministers and bankers jet in for the International Monetary Fund annual meetings a few blocks away, their delayed flights handled by a small number of unpaid air traffic controllers. There is, however, one clear message the Trump administration is notably keen to get out, not so much for its domestic audience but for the bewildered world outside. And they delivered it in the middle of last week to a small number of people ushered into the Treasury and what is said to be the finest room in Washington DC, the ornate and marbled Cash Room, which hosted the inaugural reception for post-civil war president, Ulysses Grant.

Large Language Models (LLMs) have shown remarkable capabilities in various natural language processing tasks. However, LLMs may rely on dataset biases as shortcuts for prediction, which can significantly impair their robustness and generalization capabilities. This paper presents Shortcut Suite, a comprehensive test suite designed to evaluate the impact of shortcuts on LLMs' performance, incorporating six shortcut types, five evaluation metrics, and four prompting strategies. Our extensive experiments yield several key findings: 1) LLMs demonstrate varying reliance on shortcuts for downstream tasks, significantly impairing their performance. 2) Larger LLMs are more likely to utilize shortcuts under zero-shot and few-shot in-context learning prompts. 3) Chain-of-thought prompting notably reduces shortcut reliance and outperforms other prompting strategies, while few-shot prompts generally underperform compared to zero-shot prompts. 4) LLMs often exhibit overconfidence in their predictions, especially when dealing with datasets that contain shortcuts. 5) LLMs generally have a lower explanation quality in shortcut-laden datasets, with errors falling into three types: distraction, disguised comprehension, and logical fallacy. Our findings offer new insights for evaluating robustness and generalization in LLMs and suggest potential directions for mitigating the reliance on shortcuts. The code is available at \url {https://github.com/yyhappier/ShortcutSuite.git}.

Artificial intelligence is likely to replace or at least lend a hand in tasks that take up almost three-quarters of the time bank employees now spend working. That's the conclusion of a new analysis by consultancy Accenture, which said banking has the potential to benefit more from the technology than any other industry. Just 27% of employees' time currently has a low potential of being transformed, according to the analysis. "There is a reinvention that is happening across banks, a way for firms to step back and reevaluate ways of working," Keri Smith, global banking data and AI lead at Accenture, said in an interview.

Games are a simplified model of reality and often serve as a favored platform for Artificial Intelligence (AI) research. Much of the research is concerned with game-playing agents and their decision making processes. The game of Guandan (literally, "throwing eggs") is a challenging game where even professional human players struggle to make the right decision at times. In this paper we propose a framework named GuanZero for AI agents to master this game using Monte-Carlo methods and deep neural networks. The main contribution of this paper is about regulating agents' behavior through a carefully designed neural network encoding scheme. We then demonstrate the effectiveness of the proposed framework by comparing it with state-of-the-art approaches.

We administer a Turing Test to AI Chatbots. We examine how Chatbots behave in a suite of classic behavioral games that are designed to elicit characteristics such as trust, fairness, risk-aversion, cooperation, \textit{etc.}, as well as how they respond to a traditional Big-5 psychological survey that measures personality traits. ChatGPT-4 exhibits behavioral and personality traits that are statistically indistinguishable from a random human from tens of thousands of human subjects from more than 50 countries. Chatbots also modify their behavior based on previous experience and contexts ``as if'' they were learning from the interactions, and change their behavior in response to different framings of the same strategic situation. Their behaviors are often distinct from average and modal human behaviors, in which case they tend to behave on the more altruistic and cooperative end of the distribution. We estimate that they act as if they are maximizing an average of their own and partner's payoffs.

Enterprises need to constantly look for ways to improve and expand what they offer to the marketplace. For example, Sameena Shah, managing director of AI research at JPMorgan Chase, says the company's bankers have been looking for new ways to study early-stage startups looking to raise capital. The challenge was, she says, "finding good prospects in a domain that is fundamentally very opaque and has a lot of variability." The solution for JPMorgan Chase was a new digital platform, built off an algorithm that continually seeks out data, and learns to find prospects by triaging its data into standardized representations to describe startups and likely investors. For users, the platform also offers the context of its output, to help them understand the recommendations.

A number of big themes emerged from the World Economic Forum in the Swiss resort Davos. Here are five of most pressing questions that came to dominate this year's gathering of the global elite. Donald Trump's trade war with China – continued by his successor Joe Biden – has left relations between east and west at rock bottom. But with Covid and trade tensions halving Chinese growth last year to just 3% and western businesses such as Apple moving business out of the world's second-biggest economy, Beijing has hinted it may adopt a less-hostile approach. Vice-premier Liu He appeared on the main stage at Davos to assure foreign investors that after three years of Covid disruption, it was open for business.

Card game AI has always been a hot topic in the research of artificial intelligence. In recent years, complex card games such as Mahjong, DouDizhu and Texas Hold'em have been solved and the corresponding AI programs have reached the level of human experts. In this paper, we are devoted to developing an AI program for a more complex card game, GuanDan, whose rules are similar to DouDizhu but much more complicated. To be specific, the characteristics of large state and action space, long length of one episode and the unsure number of players in the GuanDan pose great challenges for the development of the AI program. To address these issues, we propose the first AI program DanZero for GuanDan using reinforcement learning technique. Specifically, we utilize a distributed framework to train our AI system. In the actor processes, we carefully design the state features and agents generate samples by self-play. In the learner process, the model is updated by Deep Monte-Carlo Method. After training for 30 days using 160 CPUs and 1 GPU, we get our DanZero bot. We compare it with 8 baseline AI programs which are based on heuristic rules and the results reveal the outstanding performance of DanZero. We also test DanZero with human players and demonstrate its human-level performance.

AI is fantabulous and in demand in the banking and finance sector. The technological furtherance in AI – machine learning, computer vision and natural language processing has downright remodelled the business world. The expert opinion states that the growth of the AI market would reach $190 billion by the year 2025! The application of conversational assistants or chatbots is one of the substantial benefits of AI in the banking and finance sector. As opposed to an employee, a chatbot is at one's disposal 24 hours a day, and clients are more complacent using this software programme to answer inquiries and complete many typical banking procedures that traditionally called for face-to-face interaction.

The issue of data quality is foremost in the financial sector. In the financial world, abundance of data is not an issue. Data can easily be collected from a wide variety of sources such as instrument prices, news articles, stock fundamentals, social media posts, macroeconomic data, ESG data, credit card transactions, and so on. Some of this data is classified as structured and typically has a numerical quantity and a well-defined structure (e.g. stock prices). Structured data is relatively easy to feed into an ML model whereas unstructured data often requires extra processing to extract meaningful information (e.g.