Voisiko AI-tutkija olla tutkimuksen tulevaisuus?

Sakana AI Labs on aiheuttanut mullistusta tieteellisessä yhteisössä viimeisimmän ilmoituksensa myötä. Yritys on saattanut valmiiksi maailman ensimmäisen täysin autonomisen AI-tutkijan luomisen ja lanseeraamisen. Uusi algoritmi pystyy muotoilemaan teorioita, testaamaan ja paljon muuhun. Tässä on, mitä sinun tarvitsee tietää.

When you think of AI systems replacing workers in the field, there are many careers you may envision being at risk. However, most people never expected scientists to be replaced. While the latest AI scientist algorithm put forth by Sakana isn’t capable of completely replacing scientists just yet, it does deliver some unique characteristics that could leave many professionals watching its development closely.

Nykyinen tieteellisten teorioiden luomisen menetelmä

Creating scientific theories and testing them is one of the core components of research and development. Today’s scientist spent years researching questions and creating experiments to test their concepts via experimentation. Finally, the results are published and then peer-reviewed before going public. This process has been the same for over a century.
According to Sakan AI Labs’ public statements, its new AI system is capable of fully automating these tasks. The system can sift through current scientific papers, connect theories or concepts that could have undiscovered results, create new algorithms to test these theories and detail all of the experimentation for later peer review. Notably, the team focused on computational research as their primary area of focus because it enabled the experimentation to be conducted virtually.

AI:n käyttö tiedemiehillä

It’s no secret that AI has found a comfortable home within the science community. This technology can be seen integrated into various systems used by researchers today. Originally, AI systems were used to complete complex equations and simulations. Today, AI is used across the industry, where it serves even more vital roles.

One key use of AI today is sorting through all the current papers to find relevant research. Hundreds of research papers are released weekly, and the sheer volume of these releases means that the majority of game‑changing research can take years to discover. Tools like Elicit, Research Rabbit, Scite, and Consensus help scientists connect these papers in a relevant manner to build on previous data.

There are also AI tools in use today that can scan and identify specific points of interest within papers. For example, you may have a team of engineers seeking to uncover genetic mutations that cause disease. Tools like PubTator can gather similar themes to help shed light on connected research topics and experimentations.

AI-tutkijan tutkimus

These previous interactions with AI have helped the scientific community expand its horizon and have been critical in many of the latest breakthroughs. However, the introduction of a fully automated algorithm capable of researching, and publishing its papers is a major game changer. This system has the potential to uncover similarities and other hard‑to‑notice connecting factors between studies that could lead to breakthroughs in the field of computational research.

Miten AI-tutkija toimii?

The AI scientist integrates some common AI systems like Large Language Models as part of its approach. The system continually scans the troves of scientific research papers available online in repositories such as arXiv and PubMed. Here, millions of scientific research papers were used to program the LLM for a new AI algorithm. This step ensured that the system was able to speak the scientific language and could produce simalry results.

From there, the system seeks out patterns, recurring research, or connected themes. Each paper undergoes numerous checks to ensure its quality, peer‑reviewed status, time of release, and subject. Notably, the system will even cross‑reference scientific papers using sites like openreview.net. This step serves multiple roles, including ensuring quality content. In addition, this move helps to program the AI’s internal peer review system.

Toissijainen AI‑malli

Once the AI scans the repositories to find interesting experimentation and theories to test, each paper is then scored based on multiple factors. The score reflects items such as the paper’s similarity to other projects, professionalism, layout, results, and researchers. From there, the AI system examines the uniqueness of the concept and other crucial factors that help the system determine if there is additional research that should be explored or other papers that connect to a similar theory.

Tieteellisten teorioiden muotoilu

These steps enable the AI scientist to gather research and put forth connected concepts that build on the data collected. The amazing thing about the process is that it’s fully automated. No human interaction is required for this algorithm to churn out research papers. Notably, the AI focuses on computational research, which allows it to run simulations of experiments without any physical testing requirements, furthering its capabilities. As such, all tests are done using code simulations. This maneuver means that tests cost much less to complete and can be done at a higher rate.

Huolenaiheet AI‑tutkijasta

The concept of an AI scientist has raised concerns in the industry. For one, some see this approach as a lazy way to replace scientists. They argue that the system is still too new and AI is not consistent enough. Their arguments do have some basis, as there have been instances of AI inaccuracy time and time again. However, the latest algorithm includes many checks and balances, which will help to reduce these issues moving forward.

AI‑palaute silmukka

One of the biggest concerns that continues to pop up among the community is the risk of creating an AI feedback loop. This risk is very real whenever you have an AI system automatically creating, checking, and building new data models all from other AI systems. This loop can lead to errors, tunnel vision, and a disregard for human requirements. To avoid this situation, many believe a human is required in the loop.

Määrän valtava määrä

There are so many scientific papers published weekly that is nearly impossible to sort through the repositories to find the most relevant data. Currently, AI plays a vital role in helping scientists accomplish this task. However, if an AI feedback loop is created, there will be too many papers being introduced too fast, resulting in a nullification of the current AI search systems.

AI‑tutkijan hyödyt

Several benefits have many in the market excited about the prospect of an AI scientist working 24/7 searching for discoveries. The potential for this system to unlock new scientific theories is endless. As more research is added to the AI model, the system will experiment with new concepts and expand its capabilities. As such, today’s AI scientist will be nothing compared to its future iterations.

Suuret kustannussäästöt

One of the main reasons why the AI scientist is set to become a major component of the market moving forward is its major cost savings. According to the research team, their AI can create, test, publish, and peer review a scientific paper for just $15. This low cost represents a massive decrease in the cost of research and development that is impossible to deny. As such, AI scientists may become the go‑to option for those seeking to maximize their budget.

Suuri kokonaiskuva

One of the main reasons why this project has many engineers excited is that it allows researchers to connect the dots across multiple scientific studies easily. It can often seem like scientific fields develop in a bubble. However, that’s not the case. In most instances, breakthroughs in one scientific field lead to new opportunities in other fields.

AI scientists will help discover these connected theories and expand on them when possible. As such, you could see the AI scientist stacking papers to create completely new and more diverse concepts that may have slipped past human eyes. At the very least, this system will help to batch together relevant research, which will help today’s scientists see new phenomena.

Tutkijat

Sakana AI on AI‑tutkijan takana oleva tiimi. Yritys tunnetaan AI‑tieteen malleistaan. Heidän aiemmat kehityksensä sisältävät järjestelmän, joka yhdistää automaattisesti LLM:itä luodakseen kehittyneempiä järjestelmiä. He ovat myös esittäneet mekanismin, joka mahdollistaa LLM:ien käytön muiden algoritmien hienosäätöön.

Kaksi yritystä, jotka voivat hyötyä AI‑tutkijasta

The introduction of an AI scientist to the market is just the beginning of what could one day lead to a major AI scientist revolution. There are already several companies that could leverage this creation to lower their operating costs and provide better quality service to the market. Here are two firms that are poised to capitalize on this research.

1. AbCellera Biologics Inc

ABCL Hintakaavio

AbCellera Biologics Inc. tuli markkinoille vuonna 2012 pyrkien mallintamaan paremmin ihmisen vasta‑aineita tautien torjumiseksi. Yritys listautui virallisesti vuonna 2020, jolloin se sai 555 miljoonaa dollaria rahoitusta edistääkseen toimintaansa. Nykyään yritys toimii keskeisessä roolissa immuunijärjestelmän tutkimuksessa ja mallinnuksessa. Erityisesti suuri osa tästä mallinnuksesta tehdään simulointien avulla. Näin ollen AI‑tutkija soveltuu erinomaisesti näiden testien avustamiseen.

AbCellera pitää merkittävää asemaa biolääketieteen alalla. Tämän vuoksi AI‑tutkijan integrointi voisi mahdollistaa yrityksen paremman paikantamisen, testaamisen ja vertaamisen immuunivasta‑aineiden tietoihin aiemmasta tutkimuksesta. Tämä vaihe voisi paljastaa aiemmin huomaamattomia mahdollisuuksia. Nämä tekijät, yhdessä AbCellera Biologicsin asemoinnin kanssa, tekevät siitä vahvan “hold“‑sijoituksen kauppiaille.

2. Super Micro Computer

SMCI Hintakaavio

Super Micro Computer (SMCI ) perustettiin vuonna 1993 Charles Liangin ja Sara Liu:n toimesta, ja se voisi nähdä positiivisia tuloksia AI‑tutkijoiden integroinnista prosesseihinsa. Yritys on parhaiten tunnettu AI‑työkuormiin optimoiduista korkean suorituskyvyn palvelimista. Tällä hetkellä sen pääkonttori sijaitsee Kaliforniassa, ja sillä on useita valmistuslaitoksia Alankomaissa, Taiwanissa ja Yhdysvalloissa.

Super Micro Computer on yksi johtavista palvelunhallintajärjestelmien tarjoajista. Se käyttää AI:ta estämään hakkereita, valvomaan järjestelmiä ja parantamaan suorituskykyä, mikä on tehnyt yrityksestä suosikin suurille yrityksille, jotka etsivät luotettavia ja todistettuja toimittajia. Erityisesti Super Micro Computer listattiin Fortune Magazinein maailman nopeimmin kasvavien IT‑infrastruktuuriyritysten joukkoon.

Ne, jotka etsivät luotettavaa AI‑pohjaista ohjelmistoratkaisua lisättäväksi salkkuunsa, saattavat haluta vahvasti harkita SMCI:tä. Yrityksellä on vahva markkina‑asema ja se on äskettäin tehnyt sopimuksen toimittaa palvelimia Elon Muskin xAI:lle ja Teslan Gigafactorylle, joka on siirtynyt Texasiin. Kaikki nämä tekijät ovat auttaneet SMCI:tä saamaan analyytikoilta “buy“‑arvion.

AI‑tutkija avaa kustannustehokkaan tutkimuksen

Regardless of how you feel about an AI scientist pumping out research papers like comic books, many benefits make this algorithm a game changer. The sheer cost savings and future capabilities of this approach have many in the market hoping for a brighter future where discoveries occur daily while today’s scientists sleep. For now, this development is worth closely monitoring, as it’s sure to release some interesting research moving forward.

Voit oppia lisää jännittävistä AI-projekteista täältä.

(TSLA )