Top 6 AI and Machine Learning Trends for 2023
To get the full benefits of #AI and #machinelearning , IT and business leaders must develop a strategy to align AI with employee interests and business goals. The following query should be on the agenda:
Here are the top 6 #trends IT professionals get ready for in #2023.
Automated machine learning (AutoML)
Two promising aspects of #automaticmachine learning are improved #tools for data labeling and automatic optimization of neural network architectures. The demand for labeled data has created a human commentator labeling industry in low-income countries such as India, Central and Eastern Europe, and South America. The risk of using offshore labor “has forced the market to look for different ways to avoid or minimize this part of the process”. Semi-supervised and self-supervised learning enhancements help organizations minimize the amount of manually labeled data.
By automating the selection and customization of a neural network for learning models, AI becomes cheaper and brings new solutions to market in less time. Going forward, Gartner plans to focus on improving the various processes required to operationalize these models: #PlatformOps , MLOps, and #DataOps . Gartner refers to these new capabilities collectively as #XOps .
AI-enabled conceptual design
In the past, #artificialintelligence was mainly used to improve the processes of analyzing data, images, and #linguistics .?
Ideal for financial applications and in retail or healthcare and for well-defined and repetitive tasks. More recently, however, #OpenAI has developed two new models called #DALLE and #CLIP (Contrastive Language-Image Pre-training) that combine language and images to generate new visual designs from a textual description.
Early work shows how models can be trained to create new designs. Examples included an attorney's chair designed to give the AI the "Attorney's Chair" signature. New models will facilitate the implementation of production-scale AI in the creative industries.
Multi-modal learning
AI is getting better at handling multiple modalities within a single machine learning model, such as Text, image, voice, and IoT sensor data. Google DeepMind made headlines with #Gato , a multimodal AI approach capable of performing visual, linguistic, and robotic tasks. Meanwhile, developers are finding innovative ways to combine modes to improve common tasks like understanding documents.
领英推荐
For example, patient data collected and processed could include visual lab results, genetic sequence reports and clinical include data study forms, and other digitized documents. The layout and presentation style of this information can help clinicians better understand what they are seeing when done correctly. AI #algorithms trained with multimodal techniques such as computer vision and optical character recognition can optimize result presentation and improve medical diagnosis. To get the most out of multimodal techniques, data scientists must be hired or trained with cross-industry skills such as natural language processing and computer #vision techniques.
Models that can achieve multiple objectives
Typically, AI models are assigned a single goal related to a specific business metric, e.g. maximizing sales. "As early efforts mature, we can expect more companies to invest in multi-target models that address multiple goals," said Justin Silver , AI strategist and chief science officer of PROS, an intelligent sales management platform. Multi-objective models differ from multi-modal learning, which aims to learn a common representation of different types of #data .
Targeting a single business metric without considering other objectives can produce suboptimal results. For example, if a product recommendation engine only targets customer conversion rates, the company may be missing out on revenue opportunities for new or different products that the customer may not have purchased in the past. In addition, the growing importance of environmental, social, and governance (ESG) goals means that #cios need to plan models that balance sustainability goals like reducing carbon emissions and circular economy with traditional business goals like reducing inventory, delivery times, reconciling costs, etc.
AI-based cybersecurity
New #artificialintelligence and machine learning techniques will play an increasingly important role in detecting and responding to #cyberthreats . One of the main factors is that attackers are using artificial intelligence and machine learning as weapons to find vulnerabilities.
More and more companies are using AI defensively and proactively to detect unusual behavior and new attack patterns. Organizations that do not integrate #AI risk falling below the safety curve and having a higher negative impact rate.
AI-based computer programs are generally better able to manage a variety of dynamic risks cope, through both improved detection performance and enhanced agility and resilience to increased interference.
Organizations that fail to integrate AI risk falling behind the safety curve and having a higher negative impact rate.
Improved language modeling
#ChatGPT has shown new ways of thinking about incorporating AI into an interactive experience good enough for a variety of use cases across many domains, including marketing, automated customer service, and user experience.
In #2023, we expect increased demand for the quality control aspects of these AI-enhanced language models. Objections have already been raised against inaccurate coding results. For example, over the next year, companies will be confronted with inaccurate descriptions of products and dangerous recommendations. This sparks interest in finding better ways to explain how and when these tools throw errors.