Ho notlolla Geothermal Potential ka AI-Driven Software Engineering le Orchestration | B4N1

Ho notlolla Geothermal Potential ka AI-Driven Software Engineering le Orchestration

7 months ago · b4n1 Read time: 5 min

Ho notlolla Monyetla oa Geothermal ka AI-Driven Software Engineering le Orchestration

Matla a mocheso oa mocheso ke mohloli o hloekileng, o ka nchafatsoang hape o ka tšeptjoang o ka tlatsetsang haholo ho fokotsa ho tsoa ha likhase tse futhumatsang lefatše le ho fokotsa phetoho ea maemo a leholimo. Leha ho le joalo, ho sebelisa matla a tsoang mochesong oa lefatše ho hloka theknoloji e tsoetseng pele le litsamaiso tse rarahaneng ho ntša mocheso bohareng ba lefatše. Sehloohong sena, re tla hlahloba hore na boenjineri le 'mino oa li-software tse khannoang ke AI li ka notlolla matla a mocheso oa mocheso oa mocheso oa mocheso o phahameng ka ho fetesisa ka ho sebelisa AI e akaretsang ea baeloji le mefuta e meholo ea lipuo (LLMs) bakeng sa ho etsa mohlala le ho ntlafatsa lisebelisoa tsa matla a mocheso oa mocheso.

Phephetso ea Sistimi ea Matla a Geothermal

Litsamaiso tsa matla a mocheso oa mocheso li rarahane ebile lia sebetsa, li kenyelletsa mefuta e mengata e kang mocheso, khatello, sekhahla sa phallo le sebopeho sa lik'hemik'hale. Mefuta ena e sebelisana ka litsela tse sa tsitsang, e leng ho etsang hore ho be thata ho bolela esale pele le ho ntlafatsa ts'ebetso ea sistimi. Mekhoa ea setso e itšetlehile ka mefuta e matla le mekhoa e thehiloeng ho melao, eo hangata e fokolang bokhoni ba eona ba ho tšoara ho rarahana ha litsamaiso tsa mocheso oa mocheso oa mocheso.

**Ho sebelisa Generalist Biological AI le LLMs **

Li-Generalist biological AI le LLMs li fana ka mokhoa o ts'episang oa ho etsa mohlala le ho ntlafatsa lisebelisoa tsa matla a mocheso oa mocheso oa mocheso. Mefuta ena ea AI e ka ithuta ho tsoa ho li-dataset tse kholo le ho tseba mekhoa le likamano tse sa bonahaleng ka mekhoa ea setso. Ka ho sebelisa mefuta ena ea AI, re ka theha mefuta e nepahetseng le e matla ea ho bolela esale pele e ka ntlafatsang ts'ebetso ea sistimi le ho fokotsa kotsi ea ho hloleha ha lisebelisoa.

LangChain le DSPy: Lisebelisoa tsa Ochestration bakeng sa Phethahatso e Ntle

LangChain le DSPy ke lisebelisoa tse peli tse tsebahalang tsa 'mino oa liletsa tse ka thusang ho tsamaisa nts'etsopele le ho tsamaisoa ha litsamaiso tsa matla a geothermal a tsamaisoang ke AI. LangChain ke laebrari ea Python e fanang ka moralo o feto-fetohang le oa modular bakeng sa ho haha ​​​​likopo tsa AI, athe DSPy ke laebrari ea Rust e fanang ka tharollo e phahameng ea ts'ebetso le e scalable bakeng sa computing e ajoang.

Mohlala oa Code: Predictive Modelling le LangChain le LLMs

Mona ke mohlala o bonolo oa khoutu o bonts'ang mokhoa oa ho sebelisa LangChain le LLMs bakeng sa mohlala o lebelloang oa litsamaiso tsa matla a geothermal:

import langchain
ho tsoa ho langchain.llms kenya LLM


llm = LLM("davinci:3-batch")

# Hlalosa lintlha tsa sistimi ea matla a geothermal
lintlha = {
    "mocheso": [20, 30, 40, 50],
    "khatello": [10, 20, 30, 40],
    "flow_rate": [100, 200, 300, 400]
}

# Theha pompo ea LangChain
pipeline = langchain.Pipeline(
    lebitso = "geothermal_model",
    mesebetsi=[
        langchain.Tasks.TextGeneration(
            input_data=data,
            output_data={"temperature": Ha ho letho, "khatello": Ha ho letho, "flow_rate": Ha ho letho},
            llm=llm
        )
    ]
)

# Tsamaisa lipeipi
tlhahiso = pipeline.run()

# Hatisa tlhahiso
hatisa(sephetho)

Khoutu ena e hlalosa pipeline e bonolo ea LangChain e sebelisang LLM ho bolela esale pele mocheso, khatello, le tekanyo ea phallo ea tsamaiso ea matla a geothermal e thehiloeng boitsebisong ba histori.

Kaholimo ho 5-10 Litaelo/Lipotlana tse Sebelisitsoeng ka ho Fetisisa

Mona ke tse 5-10 tse sebelisoang ka ho fetesisa litaelo/likotoana tsa LangChain le DSPy:

  1. langchain.Pipeline(): E theha phaephe e ncha ea LangChain.
  2. llm=LLM("davinci:3-batch"): E hlalosa mofuta oa LLM o sebelisang mofuta oa Davinci.
  3. mesebetsi=[langchain.Tasks.TextGeneration()]: E hlalosa mosebetsi oa tlhahiso ea mongolo bakeng sa phaephe.
  4. input_data=data: E fetisa lintlha tsa nalane ho ea pele.
  5. output_data={"mocheso": Ha ho letho, "khatello": Ha ho letho, "flow_rate": Ha ho letho}: E hlalosa sebopeho sa data se hlahisoang bakeng sa phaephe.
  6. pipeline.run(): E tsamaisa phaephe mme e hlahisa tlhahiso.
  7. langchain.Tasks.TextGeneration(input_data=data, output_data=output_data, llm=llm): E hlalosa mosebetsi wa tlhahiso ya mongolo o nang le boitsebiso ba ho kenya le ho tswa.
  8. langchain.Pipeline(lebitso="geothermal_model", tasks=[task]): E theha phaephe e ncha e nang le lebitso le mosebetsi o itseng.
  9. llm.predict(input_data=data): E sebedisa LLM ho lepa sephetho ho itshetlehile ka dintlha tse kentsweng.
  10. Pipeline.update_tasks(mesebetsi=[mosebetsi_omotjha]): E ntlafatsa phaephe ka mesebetsi e metjha.

Pono ya pono

Mona ke ponelo-pele ea ts'ebetso ea ho etsa mohlala oa pele ho sebelisoa Chart.js:

{
  "type": "line",
  "data": {
    "labels": ["Mocheso", "Khatello", "Sekhahla sa Phallo"],
    "lintlha tsa boitsebiso": [
      {
        "label": "Maemo a boletsoeng esale pele",
        "data": [20, 30, 40],
        "backgroundColor": "rgba(255, 99, 132, 0.2)",
        "borderColor": "rgba(255, 99, 132, 1)",
        "pointBackgroundColor": "rgba(255, 99, 132, 1)",
        "pointBorderColor": "#fff",
        "pointHoverBackgroundColor": "#fff",
        "pointHoverBorderColor": "rgba(255, 99, 132, 1)"
      },
      {
        "label": "Litekanyetso tsa Sebele",
        "data": [25, 35, 45],
        "backgroundColor": "rgba(54, 162, 235, 0.2)",
        "borderColor": "rgba(54, 162, 235, 1)",
        "pointBackgroundColor": "rgba(54, 162, 235, 1)",
        "pointBorderColor": "#fff",
        "pointHoverBackgroundColor": "#fff",
        "pointHoverBorderColor": "rgba(54, 162, 235, 1)"
      }
    ]
  },
  "likgetho": {
    "sehlooho": {
      "bontša": 'nete,
      "text": "Predictive Modeling of Geothermal Energy Systems"
    },
    "makala": {
      "yAxes": [
        {
          "scaleLabel": {
            "bontša": 'nete,
            "labelString": "Boleng"
          }
        }
      ]
    }
  }
}

Pono ena e bonts'a litekanyetso tse boletsoeng esale pele (mohala o mofubelu) le boleng ba 'nete (mohala o moputsoa) bakeng sa mocheso, khatello le sekhahla sa phallo.

Qetello

Ho notlolla matla a mocheso oa mobu ho hloka theknoloji e tsoetseng pele le litsamaiso tse rarahaneng. Ka ho sebelisa li-Generalist biological AI le LLMs bakeng sa ho etsa mohlala le ho ntlafatsa lisebelisoa tsa matla a mocheso oa mocheso, re ka theha mefuta e nepahetseng le e matla ea ho lepa e ka ntlafatsang ts'ebetso ea sistimi le ho fokotsa kotsi ea ho se sebetse ha lisebelisoa. LangChain le DSPy ke lisebelisoa tse peli tse tsebahalang tsa 'mino oa liletsa tse ka thusang ho tsamaisa nts'etsopele le ho tsamaisoa ha litsamaiso tsa matla a geothermal a tsamaisoang ke AI. Ka ho sebelisa lisebelisoa le mekhoa ena, re ka notlolla matla a matla a mocheso oa mocheso le ho kenya letsoho bokamosong bo tsoelang pele le bo ka nchafatsoang.

Have Questions?

Let's Talk

Contact me on LinkedIn