{"675456":{"#nid":"675456","#data":{"type":"news","title":" A New Neural Network Makes Decisions Like a Human Would","body":[{"value":"\u003Cp\u003EHumans make nearly \u003Ca href=\u0022https:\/\/hbr.org\/2023\/12\/a-simple-way-to-make-better-decisions#:~:text=Various%20sources%20suggest%20that%20the,how%20we\u0026amp;apos;ll%20say%20it.\u0022\u003E35,000 decisions\u003C\/a\u003E every day, from whether it\u2019s safe to cross the road to what to have for lunch. Every decision involves weighing the options, remembering similar past scenarios, and feeling reasonably confident about the right choice. What may seem like a snap decision actually comes from gathering evidence from the surrounding environment. And often the same person makes different decisions in the same scenarios at different times.\u003C\/p\u003E\u003Cp\u003ENeural networks do the opposite, making the same decisions each time. Now, Georgia Tech researchers in Associate Professor Dobromir\u0026nbsp;Rahnev\u2019s\u0026nbsp;\u003Ca href=\u0022https:\/\/rahnevlab.gatech.edu\/\u0022\u003Elab\u003C\/a\u003E\u0026nbsp;are training them to make decisions more like humans. This science of human decision-making is only just being applied to machine learning, but developing a neural network even closer to the actual human brain may make it more reliable, according to the researchers.\u003C\/p\u003E\u003Cp\u003EIn a paper in \u003Cem\u003ENature Human Behaviour\u003C\/em\u003E, \u201c\u003Ca href=\u0022https:\/\/www.nature.com\/articles\/s41562-024-01914-8?utm_source=rct_congratemailt\u0026amp;utm_medium=email\u0026amp;utm_campaign=nonoa_20240712\u0026amp;utm_content=10.1038\/s41562-024-01914-8\u0022\u003EThe Neural Network RTNet Exhibits the Signatures of Human Perceptual Decision-Making\u003C\/a\u003E,\u201d a team from the \u003Ca href=\u0022https:\/\/psychology.gatech.edu\/\u0022\u003ESchool of Psychology\u003C\/a\u003E reveals a new neural network trained to make decisions similar to humans.\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EDecoding Decision\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003E\u201cNeural networks make a decision without telling you whether or not they are confident about their decision,\u201d said \u003Ca href=\u0022https:\/\/www.linkedin.com\/in\/farshadrafiei\/\u0022\u003EFarshad Rafiei\u003C\/a\u003E, who earned his Ph.D. in psychology at Georgia Tech. \u201cThis is one of the essential differences from how people make decisions.\u201d\u0026nbsp;\u003C\/p\u003E\u003Cp\u003ELarge language models (LLM), for example, are prone to hallucinations. When an LLM is asked a \u003Ca\u003Equestion\u003C\/a\u003E it doesn\u2019t know the answer to, it will make up something without acknowledging the artifice. By contrast, most humans in the same situation will admit they don\u2019t know the answer. Building a more human-like neural network can prevent this duplicity and lead to more accurate answers.\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EMaking the Model\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003EThe team trained their neural network on handwritten digits from a famous computer science dataset called MNIST and asked it to decipher each number. To determine the model\u2019s accuracy, they ran it with the original dataset and then added noise to the digits to make it harder for humans to discern. To compare the model performance against humans, they trained their model (as well as three other models: CNet, BLNet, and MSDNet) on the original MNIST dataset without noise, but tested them on the noisy version used in the experiments and compared results from the two datasets.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EThe researchers\u2019 model relied on two key components: a Bayesian neural network (BNN), which uses probability to make decisions, and an evidence accumulation process that keeps track of the evidence for each choice. The BNN produces responses that are slightly different each time. As it gathers more evidence, the accumulation process can sometimes favor one choice and sometimes another. Once there is enough evidence to decide, the RTNet stops the accumulation process and makes a decision.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EThe researchers also timed the model\u2019s decision-making speed to see whether it follows a psychological phenomenon called the \u201cspeed-accuracy trade-off\u201d that dictates that humans are less accurate when they must make decisions quickly.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EOnce they had the model\u2019s results, they compared them to humans\u2019 results. Sixty Georgia Tech students viewed the same dataset and shared their confidence in their decisions, and the researchers found the accuracy rate, response time, and confidence patterns were similar between the humans and the neural network.\u003C\/p\u003E\u003Cp\u003E\u201cGenerally speaking, we don\u0027t have enough human data in existing computer science literature, so we don\u0027t know how people will behave when they are exposed to these images. This limitation hinders the development of models that accurately replicate human decision-making,\u201d Rafiei said. \u201cThis work provides one of the biggest datasets of humans responding to MNIST.\u201d\u0026nbsp;\u003C\/p\u003E\u003Cp\u003ENot only did the team\u2019s model outperform all rival deterministic models, but it also was more accurate in higher-speed scenarios due to another fundamental element of human psychology: RTNet behaves like humans. As an example, people feel more confident when they make correct decisions. Without even having to train the model specifically to favor confidence, the model automatically applied it, Rafiei noted.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u201cIf we try to make our models closer to the human brain, it will show in the behavior itself without fine-tuning,\u201d he said.\u003C\/p\u003E\u003Cp\u003EThe research team hopes to train the neural network on more varied datasets to test its potential. They also expect to apply this BNN model to other neural networks to enable them to rationalize more like humans. Eventually, algorithms won\u2019t just be able to emulate our decision-making abilities, but could even help offload some of the cognitive burden of those 35,000 decisions we make daily.\u003C\/p\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003ENeural networks do the opposite, making the same decisions each time. Now, Georgia Tech researchers in Associate Professor Dobromir\u0026nbsp;Rahnev\u2019s\u0026nbsp;\u003Ca href=\u0022https:\/\/rahnevlab.gatech.edu\/\u0022\u003Elab\u003C\/a\u003E\u0026nbsp;are training them to make decisions more like humans. This science of human decision-making is only just being applied to machine learning, but developing a neural network even closer to the actual human brain may make it more reliable, according to the researchers.\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"This science of human decision-making is only just being applied to machine learning, but developing a neural network even closer to the actual human brain may make it more reliable, according to the researchers."}],"uid":"34541","created_gmt":"2024-07-15 13:52:38","changed_gmt":"2024-08-30 16:31:40","author":"Tess Malone","boilerplate_text":"","field_publication":"","field_article_url":"","dateline":{"date":"2024-07-15T00:00:00-04:00","iso_date":"2024-07-15T00:00:00-04:00","tz":"America\/New_York"},"extras":[],"groups":[{"id":"1278","name":"College of Sciences"},{"id":"1214","name":"News Room"},{"id":"1188","name":"Research Horizons"},{"id":"443951","name":"School of Psychology"}],"categories":[],"keywords":[{"id":"187915","name":"go-researchnews"},{"id":"172970","name":"go-neuro"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003ETess Malone, Senior Research Writer\/Editor\u003C\/p\u003E\u003Cp\u003Etess.malone@gatech.edu\u003C\/p\u003E","format":"limited_html"}],"email":[],"slides":[],"orientation":[],"userdata":""}}}