Hacker News

APRIL 8, 2024

This paper studies post-training large language models (LLMs) using preference feedback from a powerful oracle to help a model iteratively improve over itself. Because DNO is a batched on-policy algorithm using a regression-based objective, its implementation is straightforward and efficient.

Algorithm 89

Algorithm 89

Flipboard

APRIL 18, 2023

Two studies have shown that OpenAI's ChatGPT can decode the language used in Federal Reserve policy statements and forecast stock market trends based …

Artificial Intelligence 84

Artificial Intelligence Artificial Intelligence Machine Learning Machine Learning 84

Dataconomy

JANUARY 2, 2024

Prolific is a platform that streamlines the process for researchers to recruit participants for online studies. The platform serves researchers by providing a space to post studies and recruit suitable participants quickly and easily. Restrictions: Studies involving children as participants are not permitted on Prolific.

103

103

Hacker News

DECEMBER 25, 2023

In this work, we propose a computational model for studying a mechanism that can enable such a process. At the developmental scale, we employ these evolved reservoirs to facilitate the learning of a behavioral policy through Reinforcement Learning (RL). They do so partly due to the evolution of morphological and neural structures.

76

76

Hacker News

MARCH 2, 2024

Recent studies have found that some professionals regularly use cannabis while programming even for work-related tasks. This paper presents the first controlled observational study of the effects of cannabis on programming ability. Anecdotal evidence of cannabis use by professional programmers abounds.

73

73

Data Science Dojo

OCTOBER 10, 2023

Learning from real-world applications : Who doesn’t want to revolutionize their manufacturing process by integrating AI, a strategy learned from a case study at an AI conference. Speakers included experts on AI ethics, law, and policy, as well as representatives from industry and academia.

AI 243

AI AI Artificial Intelligence Artificial Intelligence 243

Hacker News

FEBRUARY 7, 2024

Moreover, responses in these datasets are often sampled from a language model distinct from the one being aligned, and since the model evolves over training, the alignment phase is inevitably off-policy. In this study, we posit that online feedback is key and improves DAP methods.

AI 87

AI AI 87

Dataconomy

MAY 1, 2024

Amazon Q also makes it easier for employees to get answers to questions across business data, such as company policies, product information, business results, code base, employees, and many other topics by connecting to enterprise data repositories to summarize the data logically, analyze trends, and engage in dialog about the data.”

AWS 113

AWS AI AI Cloud Computing 113

Dataconomy

MARCH 19, 2024

In a recent study by Melita , it was made clear that ESG is no longer just a tick-box exercise. Regulatory policy trends Regulators play a vital role in advancing sustainability standards in telecom. Such policy measures are creating an imperative for telecom companies to take tangible climate action.

AI 192

AI AI 192

Data Science Blog

APRIL 25, 2023

This is the second part of the fourth article of the series My elaborate study notes on reinforcement learning. *In 10, N-step returns and their weighted sum If you remember differences between approximations of the value function for a policy by Monte Carlo methods or TD methods, this section is nothing hard to read.

Algorithm 147

Algorithm Deep Learning Deep Learning Machine Learning 147

Hacker News

MARCH 28, 2024

On Thursday, the climate-data sleuths at Carbon Mapper published a study in Science that shows the nation’s landfills emit that gas at levels at least 40 percent higher than previously reported to the Environmental Protection Agency. We need to have regulations in place.”

176

176

DrivenData Labs

MARCH 26, 2024

Honorable Mentions gallery Meet the winners ¶ Best Overall Prize Five teams will have the opportunity to attend a 10-day Space Study program with travel, lodging, and tuition covered. Sebastian Sanches Bernal is studying Remote Sensing and Geospatial Sciences at Boston University. What motivated you to choose this topic?

Power BI 264

Power BI Data Scientist Python Machine Learning 264

Dataconomy

SEPTEMBER 11, 2023

A recent study by the Corporate Europe Organisation (CEO) pulled back the curtain, revealing how much these tech behemoths are investing in lobbying efforts to shape EU policies. The influence of tech giants extends far beyond our gadgets and apps. surge since 2021, when their lobbying budget stood at €97 million.

Artificial Intelligence 103

Artificial Intelligence Artificial Intelligence AI AI 103

IBM Journey to AI blog

OCTOBER 24, 2023

” The case study in this blog post shows how to easily and securely implement and manage a site reliability engineering (SRE) team’s access across an enterprise. Case study A large banking client has a centralized site reliability engineering (SRE) team that manages operations for all resources in the organization. .”

Database 68

Database Clustering 68

Dataconomy

MAY 29, 2024

While you focus on your college education, studying hard and completing your assignments, behind the scenes, your university’s IT department is battling to protect both your information and the institution’s data from cyber threats. A unified security policy for all devices is critical to manage this challenge effectively.

113

113

Dataconomy

NOVEMBER 7, 2023

Image credit ) Usage policies and data limits After checking the availability of all other features, one important feature is the company’s policies and data limits… A company might add policies that you find out later and don’t suit your convenience, it’s always better to check these early on to avoid the issues later on.

103

103

IBM Journey to AI blog

DECEMBER 8, 2023

A recent report from the IBM Institute for Business Value found that 68% of hybrid cloud users already have a formal, organization-wide policy or approach for the use of generative AI. Our intensive 3-day Study Jams, with presentations and interactive study for 2 hours each day, will focus on hybrid cloud and generative AI during 2H 2024.

Cloud Computing 68

Cloud Computing AI AI 68

IBM Journey to AI blog

DECEMBER 4, 2023

In a new case study featuring IBM, Gartner talks about how to establish a governance framework to streamline the process of detecting and managing technology ethics concerns in AI projects. GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S.

AI 78

AI AI Artificial Intelligence Artificial Intelligence 78

Smart Data Collective

OCTOBER 8, 2022

Some of them are even using the cloud to improve their studying capabilities. It requires ample preparation, studying, and practice. However, since the scoring format changed in 2016, schools may tweak their policies accordingly. When studying for the SAT, you’ll want to create a schedule. Create a Schedule.

Cloud Computing 129

Cloud Computing Big Data Big Data AI 129

NYU Center for Data Science

MARCH 23, 2023

The AI, Misinformation, and Policy Seminar Series (AMPol) at the Center for Data Science explores this critical research area, featuring speakers working in the intersecting fields of data science, machine learning, and misinformation. To access the lecture slides, please visit Emily Saltz Lecture Slides. by Meryl Phair

Data Science 71

Data Science AI AI Artificial Intelligence 71

BAIR

OCTOBER 16, 2023

By incorporating a new component - pairwise policy gradient, we can unify the reward modeling stage and RL stage, enabling direct updates based on pairwise responses. Proximal Policy Optimization ( PPO ), the dominant RL optimizer in this process, has been reported to exhibit instability and implementation complications.

Algorithm 90

Algorithm 90

Dataconomy

OCTOBER 16, 2023

Cramly AI is your newfound study buddy! Dive deeper into your studies, and let Cramly light the way! Ever felt the need for an instant study buddy? Doctrina AI can easily be your best study ally How to use Cramly AI? One cannot deny the immense impact of the Cold War on international relations and domestic policies alike.

AI 187

AI AI Natural Language Processing Artificial Intelligence 187

ML @ CMU

DECEMBER 1, 2023

We study peer reviewing of peer reviews driven by two primary motivations: (i) Incentivizing reviewers to provide high-quality reviews is an important open problem. ii) Many experiments in the peer-review processes of various scientific fields use evaluations of reviews as a “gold standard” for investigating policies and interventions.

Machine Learning 159

Machine Learning Machine Learning 159

NYU Center for Data Science

JUNE 16, 2023

Motivated by applications in healthcare and criminal justice, Umang studies how to create algorithmic decision-making systems endowed with the ability to explain their behavior and adapt to a stakeholder’s expertise to improve human-machine team performance. “My IEEE Spectrum , Amazon Science ) and referenced in policy briefs (e.g.,

Machine Learning 108

Machine Learning Machine Learning Artificial Intelligence Artificial Intelligence 108

BAIR

JULY 10, 2022

In cooperative multi-agent reinforcement learning (MARL), due to its on-policy nature, policy gradient (PG) methods are typically believed to be less sample efficient than value decomposition (VD) methods, which are off-policy. By contrast, PG methods with individual policies can converge to an optimal policy in these cases.

Algorithm 59

Algorithm 59

Hacker News

MARCH 28, 2023

The result is a 12-part series and e-book, The EV Transition Explained , that explores the tightly woven technological, policy, and social issues involved. I also reviewed hundreds of reports, case studies, and books surrounding EVs and electrical grids. Consider the reactions to the recent change in U.S.

163

163

NYU Center for Data Science

MAY 17, 2024

I always wanted to do a PhD in physics and study dark matter and dark energy to understand the rest of the universe. But during COVID, I had an epiphany: I cannot study the sky if my house is on fire. “I was suddenly in my bedroom finishing my last year in college,” he recalled. “I I found it too distracting.” By Stephen Thomas

Machine Learning 71

Machine Learning Machine Learning Algorithm Data Science 71

ML @ CMU

FEBRUARY 24, 2023

We build our policy network by training a task-specific multi-layer perceptron (MLP) network inserted into a frozen pre-trained LM. Discrete prompt optimization thus amounts to learning a small number of policy parameters which we set as an MLP layer inserted into a frozen compact model such as distilGPT-2. of our paper. of our paper.

Natural Language Processing 253

Natural Language Processing Algorithm Computer Science Computer Science 253

IBM Journey to AI blog

JULY 24, 2023

By adding time-based restrictions , you can scope these access policies further to a specific time and date range (once) or to recurring windows. When creating a new policy, you can now optionally add conditions for when the access should be granted. Once again, I deploy everything with Terraform, including IAM privileges.

81

81

IBM Journey to AI blog

MAY 31, 2024

However, the latest CEO Study by the IBM Institute for the Business Value found that 72% of the surveyed government leaders say that the potential productivity gains from AI and automation are so great that they must accept significant risk to stay competitive. What’s next?

AI 74

AI AI Artificial Intelligence Artificial Intelligence 74

Chatbots Life

SEPTEMBER 11, 2023

Suggested Read: How to add dialogflow user name reply Case study scenario We can consider an Insurance chatbot example using follow-up intents. Here we have set a default welcome with the Insurance Policy types available with a button option “Check our Insurance Policies”. These three policies are the follow-up intents.

AI 52

AI AI 52

AssemblyAI

DECEMBER 23, 2022

Reinforcement Learning from Human Feedback The method overall consists of three distinct steps: Supervised fine-tuning step: a pre-trained language model is fine-tuned on a relatively small amount of demonstration data curated by labelers, to learn a supervised policy (the SFT model ) that generates outputs from a selected list of prompts.

Supervised Learning 98

Supervised Learning Algorithm Deep Learning Deep Learning 98

Precisely

FEBRUARY 23, 2024

Even before the age of advanced analytics, experts in the industry were routinely using data to assess risk and price policies. Of the senior insurance executives and managers who participated in the study, nearly half said that their organization’s data-driven digital transformation efforts match those of other carriers.

Data Quality 62

Data Quality Data Governance Analytics Analytics 62

Mlearning.ai

JULY 14, 2023

This time, I will delve deeper into the utilisation of Proximal Policy Optimisation (PPO), a reinforcement learning approach, in the training process of ChatGPT. In this article, I will provide an introduction to reinforcement learning, starting from the basics and covering both value-based and policy-based RL methods.

Algorithm 52

Algorithm Deep Learning Deep Learning Machine Learning 52

Data Science Blog

APRIL 13, 2023

In a 2022 study, 86% of patients agreed their health-monitoring wearables improved their health and quality of life. These insights, in turn, guide more effective decision-making on reorganizing workflows or editing policies to improve standards of living or reduce costs.

Internet of Things 147

Internet of Things Analytics Analytics Big Data 147

Flipboard

DECEMBER 27, 2023

Not everyone is against surveillance,” says anthropology graduate student Matheus Viegas Ferrari of the Universidade Federal da Bahia, in Brazil, and the Université Paris 8: Saint-Denis, in Paris, who attended the community meeting in Seine-Saint-Denis and published a study of surveillance at the 2024 Olympics.

Algorithm 158

Algorithm Internet of Things AI AI 158

Dataconomy

OCTOBER 11, 2023

This move provided several options, such as computer science studies and involvement in competitive math programs. Although she has postponed her college plans in order to focus on her growing business, Pranjali is open to furthering her studies in the future in order to get essential business skills.

AI 185

AI AI Data Silos Computer Science 185