মেশিন লার্নিং শব্দকোষ: ML ফান্ডামেন্টালস

এই পৃষ্ঠায় ML ফান্ডামেন্টাল শব্দকোষের পদ রয়েছে। সকল শব্দকোষের জন্য এখানে ক্লিক করুন

সঠিকতা

#মৌলিক

সঠিক শ্রেণীবিভাগের ভবিষ্যদ্বাণীর সংখ্যাকে ভবিষ্যদ্বাণীর মোট সংখ্যা দিয়ে ভাগ করলে। এটাই:

$$\text{Accuracy} = \frac{\text{correct predictions}} {\text{correct predictions + incorrect predictions }}$$

উদাহরণস্বরূপ, একটি মডেল যা 40টি সঠিক ভবিষ্যদ্বাণী করেছে এবং 10টি ভুল ভবিষ্যদ্বাণী করেছে তার সঠিকতা থাকবে:

$$\text{Accuracy} = \frac{\text{40}} {\text{40 + 10}} = \text{80%}$$

বাইনারি শ্রেণীবিভাগ সঠিক ভবিষ্যদ্বাণী এবং ভুল ভবিষ্যদ্বাণীর বিভিন্ন বিভাগের জন্য নির্দিষ্ট নাম প্রদান করে। সুতরাং, বাইনারি শ্রেণীবিভাগের নির্ভুলতা সূত্রটি নিম্নরূপ:

$$\text{Accuracy} = \frac{\text{TP} + \text{TN}} {\text{TP} + \text{TN} + \text{FP} + \text{FN}}$$

কোথায়:

নির্ভুলতা এবং প্রত্যাহার সঙ্গে তুলনা এবং বিপরীতে নির্ভুলতা.

সক্রিয়করণ ফাংশন

#মৌলিক

একটি ফাংশন যা নিউরাল নেটওয়ার্কগুলিকে বৈশিষ্ট্য এবং লেবেলের মধ্যে অরৈখিক (জটিল) সম্পর্ক শিখতে সক্ষম করে।

জনপ্রিয় সক্রিয়করণ ফাংশন অন্তর্ভুক্ত:

অ্যাক্টিভেশন ফাংশনগুলির প্লটগুলি কখনই একক সরলরেখা নয়। উদাহরণস্বরূপ, ReLU অ্যাক্টিভেশন ফাংশনের প্লট দুটি সরল রেখা নিয়ে গঠিত:

দুটি লাইনের একটি কার্টেসিয়ান প্লট। প্রথম লাইনের একটি ধ্রুবক y মান আছে 0, x-অক্ষ বরাবর -ইনফিনিটি,0 থেকে 0,-0 পর্যন্ত চলছে। দ্বিতীয় লাইন 0,0 এ শুরু হয়। এই লাইনে +1 এর ঢাল রয়েছে, তাই এটি 0,0 থেকে +ইনফিনিটি,+ইনফিনিটি পর্যন্ত চলে।

সিগমায়েড অ্যাক্টিভেশন ফাংশনের একটি প্লট নিম্নরূপ দেখায়:

একটি দ্বি-মাত্রিক বাঁকা প্লট যেখানে x মানগুলি ডোমেইন -ইনফিনিটি থেকে + পজিটিভ পর্যন্ত বিস্তৃত, যখন y মানগুলি 0 থেকে প্রায় 1 পর্যন্ত বিস্তৃত। যখন x 0 হয়, y হয় 0.5। বক্ররেখার ঢাল সর্বদাই ধনাত্মক, সর্বোচ্চ ঢাল 0,0.5 এবং ধীরে ধীরে ঢাল কমতে থাকে কারণ x এর পরম মান বৃদ্ধি পায়।

কৃত্রিম বুদ্ধিমত্তা

#মৌলিক

একটি অ-মানব প্রোগ্রাম বা মডেল যা পরিশীলিত কাজগুলি সমাধান করতে পারে। উদাহরণস্বরূপ, একটি প্রোগ্রাম বা মডেল যা পাঠ্য অনুবাদ করে বা একটি প্রোগ্রাম বা মডেল যা রেডিওলজিক চিত্র থেকে রোগ সনাক্ত করে উভয়ই কৃত্রিম বুদ্ধিমত্তা প্রদর্শন করে।

আনুষ্ঠানিকভাবে, মেশিন লার্নিং হল কৃত্রিম বুদ্ধিমত্তার একটি উপ-ক্ষেত্র। যাইহোক, সাম্প্রতিক বছরগুলিতে, কিছু সংস্থা কৃত্রিম বুদ্ধিমত্তা এবং মেশিন লার্নিং শব্দগুলিকে বিনিময়যোগ্যভাবে ব্যবহার করা শুরু করেছে।

AUC (ROC বক্ররেখার অধীনে এলাকা)

#মৌলিক

0.0 এবং 1.0 এর মধ্যে একটি সংখ্যা যা একটি বাইনারি শ্রেণীবিন্যাস মডেলের নেতিবাচক শ্রেণী থেকে ইতিবাচক শ্রেণীগুলিকে পৃথক করার ক্ষমতা উপস্থাপন করে। AUC 1.0 এর যত কাছাকাছি হবে, মডেলের একে অপরের থেকে ক্লাস আলাদা করার ক্ষমতা তত ভাল।

উদাহরণস্বরূপ, নিম্নলিখিত চিত্রটি একটি শ্রেণীবদ্ধ মডেল দেখায় যা ইতিবাচক শ্রেণীগুলি (সবুজ ডিম্বাকৃতি) নেতিবাচক শ্রেণীগুলি (বেগুনি আয়তক্ষেত্রগুলি) থেকে পুরোপুরি আলাদা করে। এই অবাস্তবভাবে নিখুঁত মডেলটির একটি AUC 1.0 রয়েছে:

এক পাশে 8টি ইতিবাচক উদাহরণ এবং অন্য পাশে 9টি নেতিবাচক উদাহরণ সহ একটি সংখ্যা রেখা৷

বিপরীতভাবে, নিম্নলিখিত চিত্রটি একটি শ্রেণিবদ্ধ মডেলের ফলাফল দেখায় যা এলোমেলো ফলাফল তৈরি করেছে। এই মডেলটির একটি AUC 0.5 রয়েছে:

6টি ইতিবাচক উদাহরণ এবং 6টি নেতিবাচক উদাহরণ সহ একটি সংখ্যা রেখা৷ উদাহরণের ক্রম হল ধনাত্মক, ঋণাত্মক, ধনাত্মক, ধনাত্মক, ধনাত্মক, নেতিবাচক, ধনাত্মক, নেতিবাচক, ধনাত্মক, ধনাত্মক, নেতিবাচক।

হ্যাঁ, পূর্ববর্তী মডেলটির একটি AUC 0.5 আছে, 0.0 নয়৷

বেশিরভাগ মডেল দুটি চরমের মধ্যে কোথাও আছে। উদাহরণস্বরূপ, নিম্নলিখিত মডেলটি নেতিবাচক থেকে ইতিবাচককে কিছুটা আলাদা করে, এবং তাই 0.5 এবং 1.0 এর মধ্যে একটি AUC রয়েছে:

6টি ইতিবাচক উদাহরণ এবং 6টি নেতিবাচক উদাহরণ সহ একটি সংখ্যা রেখা৷ উদাহরণের ক্রম হল ঋণাত্মক, ঋণাত্মক, ঋণাত্মক, ঋণাত্মক, নেতিবাচক, নেতিবাচক, ধনাত্মক, ধনাত্মক, ধনাত্মক, ধনাত্মক, ধনাত্মক।

AUC আপনার শ্রেণীবিন্যাস থ্রেশহোল্ডের জন্য সেট করা যেকোনো মান উপেক্ষা করে। পরিবর্তে, AUC সমস্ত সম্ভাব্য শ্রেণীবিভাগ থ্রেশহোল্ড বিবেচনা করে।

backpropagation

#মৌলিক

অ্যালগরিদম যা নিউরাল নেটওয়ার্কগুলিতে গ্রেডিয়েন্ট ডিসেন্ট প্রয়োগ করে৷

একটি নিউরাল নেটওয়ার্ক প্রশিক্ষণ নিম্নলিখিত দুই-পাস চক্রের অনেক পুনরাবৃত্তি জড়িত:

  1. ফরওয়ার্ড পাসের সময়, সিস্টেম ভবিষ্যদ্বাণী(গুলি) প্রদানের জন্য উদাহরণগুলির একটি ব্যাচ প্রক্রিয়া করে। সিস্টেম প্রতিটি লেবেল মানের সাথে প্রতিটি ভবিষ্যদ্বাণী তুলনা করে। ভবিষ্যদ্বাণী এবং লেবেল মানের মধ্যে পার্থক্য হল সেই উদাহরণের ক্ষতি । বর্তমান ব্যাচের মোট ক্ষতি গণনা করার জন্য সিস্টেমটি সমস্ত উদাহরণের জন্য ক্ষতিগুলিকে একত্রিত করে।
  2. ব্যাকওয়ার্ড পাসের সময় (ব্যাকপ্রপাগেশন), সিস্টেমটি সমস্ত লুকানো স্তর(গুলি) এর সমস্ত নিউরনের ওজন সামঞ্জস্য করে ক্ষতি হ্রাস করে।

নিউরাল নেটওয়ার্কে প্রায়ই অনেক লুকানো স্তর জুড়ে অনেক নিউরন থাকে। এই নিউরনগুলির প্রতিটি বিভিন্ন উপায়ে সামগ্রিক ক্ষতিতে অবদান রাখে। ব্যাকপ্রোপ্যাগেশন নির্দিষ্ট নিউরনগুলিতে প্রয়োগ করা ওজন বাড়ানো বা হ্রাস করা নির্ধারণ করে।

শেখার হার হল একটি গুণক যা প্রতিটি পশ্চাদগামী পাস প্রতিটি ওজন বৃদ্ধি বা হ্রাস করার মাত্রা নিয়ন্ত্রণ করে। একটি বড় শেখার হার প্রতিটি ওজন একটি ছোট শেখার হারের চেয়ে বেশি বৃদ্ধি বা হ্রাস করবে।

ক্যালকুলাস পরিভাষায়, ব্যাকপ্রপাগেশন ক্যালকুলাসের চেইন নিয়ম প্রয়োগ করে। অর্থাৎ, ব্যাকপ্রোপগেশন প্রতিটি প্যারামিটারের ক্ষেত্রে ত্রুটির আংশিক ডেরিভেটিভ গণনা করে। আরো বিস্তারিত জানার জন্য, মেশিন লার্নিং ক্র্যাশ কোর্সের এই টিউটোরিয়ালটি দেখুন।

কয়েক বছর আগে, এমএল অনুশীলনকারীদের ব্যাকপ্রোপগেশন বাস্তবায়নের জন্য কোড লিখতে হয়েছিল। TensorFlow-এর মতো আধুনিক ML APIগুলি এখন আপনার জন্য ব্যাকপ্রোপগেশন প্রয়োগ করে৷ ফাউ!

ব্যাচ

#মৌলিক

একটি প্রশিক্ষণের পুনরাবৃত্তিতে ব্যবহৃত উদাহরণের সেট। ব্যাচের আকার একটি ব্যাচে উদাহরণের সংখ্যা নির্ধারণ করে।

একটি ব্যাচ কিভাবে একটি যুগের সাথে সম্পর্কিত তার ব্যাখ্যার জন্য epoch দেখুন।

ব্যাচ আকার

#মৌলিক

একটি ব্যাচে উদাহরণের সংখ্যা। উদাহরণস্বরূপ, যদি ব্যাচের আকার 100 হয়, তাহলে মডেলটি প্রতি পুনরাবৃত্তি 100টি উদাহরণ প্রক্রিয়া করে।

নিম্নলিখিত জনপ্রিয় ব্যাচ আকার কৌশল:

  • স্টোকাস্টিক গ্রেডিয়েন্ট ডিসেন্ট (SGD) , যার ব্যাচের আকার 1।
  • সম্পূর্ণ ব্যাচ, যেখানে ব্যাচের আকার পুরো প্রশিক্ষণ সেটের উদাহরণের সংখ্যা। উদাহরণস্বরূপ, যদি প্রশিক্ষণ সেটে এক মিলিয়ন উদাহরণ থাকে, তাহলে ব্যাচের আকার এক মিলিয়ন উদাহরণ হবে। সম্পূর্ণ ব্যাচ সাধারণত একটি অদক্ষ কৌশল।
  • মিনি-ব্যাচ যেখানে ব্যাচের আকার সাধারণত 10 থেকে 1000 এর মধ্যে হয়। মিনি-ব্যাচ সাধারণত সবচেয়ে কার্যকরী কৌশল।

পক্ষপাত (নৈতিকতা/ন্যায়)

#ন্যায্যতা
#মৌলিক

1. কিছু জিনিস, মানুষ বা গোষ্ঠীর প্রতি স্টিরিওটাইপিং, কুসংস্কার বা পক্ষপাতিত্ব। এই পক্ষপাতগুলি ডেটা সংগ্রহ এবং ব্যাখ্যা, একটি সিস্টেমের নকশা এবং ব্যবহারকারীরা কীভাবে একটি সিস্টেমের সাথে যোগাযোগ করে তা প্রভাবিত করতে পারে। এই ধরনের পক্ষপাতের ফর্মগুলির মধ্যে রয়েছে:

2. একটি নমুনা বা রিপোর্টিং পদ্ধতি দ্বারা প্রবর্তিত পদ্ধতিগত ত্রুটি। এই ধরনের পক্ষপাতের ফর্মগুলির মধ্যে রয়েছে:

মেশিন লার্নিং মডেল বা ভবিষ্যদ্বাণী পক্ষপাতিত্ব শব্দের সাথে বিভ্রান্ত হবেন না।

পক্ষপাত (গণিত) বা পক্ষপাত শব্দ

#মৌলিক

একটি উত্স থেকে একটি বাধা বা অফসেট. বায়াস হল মেশিন লার্নিং মডেলের একটি প্যারামিটার, যা নিম্নলিখিতগুলির যে কোনো একটি দ্বারা চিহ্নিত করা হয়:

  • w 0

উদাহরণস্বরূপ, নিম্নোক্ত সূত্রে পক্ষপাত হল b :

$$y' = b + w_1x_1 + w_2x_2 + … w_nx_n$$

একটি সাধারণ দ্বি-মাত্রিক লাইনে, পক্ষপাত মানে শুধু "y-ইন্টারসেপ্ট।" উদাহরণস্বরূপ, নিম্নলিখিত চিত্রে লাইনের পক্ষপাত হল 2।

0.5 এর ঢাল সহ একটি লাইনের প্লট এবং 2 এর একটি পক্ষপাত (y-ইন্টারসেপ্ট)।

পক্ষপাত বিদ্যমান কারণ সমস্ত মডেল মূল (0,0) থেকে শুরু হয় না। উদাহরণস্বরূপ, ধরুন একটি বিনোদন পার্কে প্রবেশের জন্য 2 ইউরো এবং একজন গ্রাহক থাকার প্রতি ঘন্টার জন্য অতিরিক্ত 0.5 ইউরো। অতএব, মোট খরচ ম্যাপিং একটি মডেল 2 এর পক্ষপাতী কারণ সর্বনিম্ন খরচ হল 2 ইউরো।

নৈতিকতা এবং ন্যায্যতা বা ভবিষ্যদ্বাণী পক্ষপাতের সাথে পক্ষপাতিত্বের সাথে বিভ্রান্ত হওয়া উচিত নয়।

বাইনারি শ্রেণীবিভাগ

#মৌলিক

এক ধরনের শ্রেণীবিন্যাস টাস্ক যা দুটি পারস্পরিক একচেটিয়া শ্রেণীর একটির পূর্বাভাস দেয়:

উদাহরণস্বরূপ, নিম্নলিখিত দুটি মেশিন লার্নিং মডেল প্রতিটি বাইনারি শ্রেণীবিভাগ সম্পাদন করে:

  • একটি মডেল যা নির্ধারণ করে যে ইমেল বার্তাগুলি স্প্যাম (পজিটিভ ক্লাস) না স্প্যাম (নেতিবাচক ক্লাস)।
  • একটি মডেল যা একজন ব্যক্তির একটি নির্দিষ্ট রোগ (পজিটিভ ক্লাস) আছে কিনা বা সেই রোগ (নেতিবাচক শ্রেণী) নেই কিনা তা নির্ধারণ করতে চিকিৎসা লক্ষণগুলি মূল্যায়ন করে।

বহু-শ্রেণীর শ্রেণীবিভাগের সাথে বৈসাদৃশ্য।

লজিস্টিক রিগ্রেশন এবং শ্রেণীবিভাগ থ্রেশহোল্ড দেখুন।

বালতি

#মৌলিক

একটি একক বৈশিষ্ট্যকে একাধিক বাইনারি বৈশিষ্ট্যে রূপান্তর করা যাকে বলা হয় বালতি বা বিন , সাধারণত একটি মান পরিসরের উপর ভিত্তি করে। কাটা বৈশিষ্ট্যটি সাধারণত একটি অবিচ্ছিন্ন বৈশিষ্ট্য

উদাহরণস্বরূপ, তাপমাত্রাকে একটি অবিচ্ছিন্ন ফ্লোটিং-পয়েন্ট বৈশিষ্ট্য হিসাবে উপস্থাপন করার পরিবর্তে, আপনি তাপমাত্রার রেঞ্জগুলিকে আলাদা বালতিতে কাটতে পারেন, যেমন:

  • <= 10 ডিগ্রি সেলসিয়াস হবে "ঠান্ডা" বালতি।
  • 11 - 24 ডিগ্রি সেলসিয়াস "নাতিশীতোষ্ণ" বালতি হবে।
  • >= 25 ডিগ্রি সেলসিয়াস হবে "উষ্ণ" বালতি।

মডেলটি একই বালতিতে প্রতিটি মানকে অভিন্নভাবে বিবেচনা করবে। উদাহরণস্বরূপ, মান 13 এবং 22 উভয়ই নাতিশীতোষ্ণ বালতিতে রয়েছে, তাই মডেল দুটি মানকে অভিন্নভাবে বিবেচনা করে।

শ্রেণীবদ্ধ তথ্য

#মৌলিক

সম্ভাব্য মানগুলির একটি নির্দিষ্ট সেট থাকা বৈশিষ্ট্যগুলি ৷ উদাহরণ স্বরূপ, traffic-light-state নামে একটি শ্রেণীবদ্ধ বৈশিষ্ট্য বিবেচনা করুন, যেটিতে শুধুমাত্র নিম্নলিখিত তিনটি সম্ভাব্য মানগুলির মধ্যে একটি থাকতে পারে:

  • red
  • yellow
  • green

traffic-light-state একটি শ্রেণীবদ্ধ বৈশিষ্ট্য হিসাবে উপস্থাপন করে, একটি মডেল চালকের আচরণের উপর red , green এবং yellow বিভিন্ন প্রভাব শিখতে পারে।

শ্রেণীগত বৈশিষ্ট্যগুলিকে কখনও কখনও পৃথক বৈশিষ্ট্য বলা হয়।

সংখ্যাসূচক তথ্যের সাথে বৈসাদৃশ্য।

ক্লাস

#মৌলিক

একটি বিভাগ যা একটি লেবেল অন্তর্গত হতে পারে৷ উদাহরণ স্বরূপ:

একটি শ্রেণিবিন্যাস মডেল একটি শ্রেণির পূর্বাভাস দেয়। বিপরীতে, একটি রিগ্রেশন মডেল একটি শ্রেণির পরিবর্তে একটি সংখ্যার পূর্বাভাস দেয়।

শ্রেণিবিন্যাস মডেল

#মৌলিক

একটি মডেল যার ভবিষ্যদ্বাণী একটি ক্লাস । উদাহরণস্বরূপ, নিম্নলিখিত সমস্ত শ্রেণীবিভাগ মডেল:

  • একটি মডেল যা একটি ইনপুট বাক্যের ভাষার ভবিষ্যদ্বাণী করে (ফরাসি? স্প্যানিশ? ইতালীয়?)।
  • একটি মডেল যা গাছের প্রজাতির ভবিষ্যদ্বাণী করে (ম্যাপেল? ওক? বাওবাব?)।
  • একটি মডেল যা একটি নির্দিষ্ট চিকিৎসা অবস্থার জন্য ইতিবাচক বা নেতিবাচক শ্রেণীর পূর্বাভাস দেয়।

বিপরীতে, রিগ্রেশন মডেলগুলি ক্লাসের পরিবর্তে সংখ্যার পূর্বাভাস দেয়।

দুটি সাধারণ ধরনের শ্রেণিবিন্যাস মডেল হল:

শ্রেণীবিভাগ থ্রেশহোল্ড

#মৌলিক

একটি বাইনারি শ্রেণীবিভাগে , 0 এবং 1 এর মধ্যে একটি সংখ্যা যা একটি লজিস্টিক রিগ্রেশন মডেলের কাঁচা আউটপুটকে ধনাত্মক শ্রেণি বা নেতিবাচক শ্রেণির একটি ভবিষ্যদ্বাণীতে রূপান্তর করে। মনে রাখবেন যে শ্রেণীবিন্যাস থ্রেশহোল্ড এমন একটি মান যা একজন মানুষ বেছে নেয়, মডেল প্রশিক্ষণ দ্বারা নির্বাচিত একটি মান নয়।

একটি লজিস্টিক রিগ্রেশন মডেল 0 এবং 1 এর মধ্যে একটি কাঁচা মান আউটপুট করে। তারপর:

  • যদি এই কাঁচা মানটি শ্রেণিবিন্যাসের থ্রেশহোল্ডের চেয়ে বেশি হয়, তাহলে ধনাত্মক শ্রেণির পূর্বাভাস দেওয়া হয়।
  • যদি এই কাঁচা মানটি শ্রেণিবিন্যাসের থ্রেশহোল্ডের চেয়ে কম হয়, তাহলে নেতিবাচক শ্রেণির পূর্বাভাস দেওয়া হয়।

উদাহরণস্বরূপ, ধরুন শ্রেণীবিন্যাস থ্রেশহোল্ড হল 0.8। যদি কাঁচা মান 0.9 হয়, তাহলে মডেলটি ইতিবাচক শ্রেণীর পূর্বাভাস দেয়। যদি কাঁচা মান 0.7 হয়, তাহলে মডেলটি নেতিবাচক শ্রেণীর পূর্বাভাস দেয়।

শ্রেণীবিন্যাস থ্রেশহোল্ডের পছন্দ দৃঢ়ভাবে মিথ্যা ইতিবাচক এবং মিথ্যা নেতিবাচক সংখ্যাকে প্রভাবিত করে।

শ্রেণী-ভারসাম্যহীন ডেটাসেট

#মৌলিক

একটি শ্রেণিবিন্যাসের সমস্যার জন্য একটি ডেটাসেট যেখানে প্রতিটি শ্রেণীর লেবেলের মোট সংখ্যা উল্লেখযোগ্যভাবে আলাদা। উদাহরণস্বরূপ, একটি বাইনারি শ্রেণিবিন্যাস ডেটাসেট বিবেচনা করুন যার দুটি লেবেল নিম্নরূপ বিভক্ত:

  • 1,000,000 নেতিবাচক লেবেল
  • 10টি ইতিবাচক লেবেল

নেতিবাচক থেকে ইতিবাচক লেবেলের অনুপাত হল 100,000 থেকে 1, তাই এটি একটি শ্রেণী-ভারসাম্যহীন ডেটাসেট।

বিপরীতে, নিম্নলিখিত ডেটাসেটটি শ্রেণী-ভারসাম্যহীন নয় কারণ নেতিবাচক লেবেলের সাথে ইতিবাচক লেবেলের অনুপাত তুলনামূলকভাবে 1-এর কাছাকাছি:

  • 517 নেতিবাচক লেবেল
  • 483 ইতিবাচক লেবেল

মাল্টি-ক্লাস ডেটাসেটগুলিও শ্রেণী-ভারসাম্যহীন হতে পারে। উদাহরণস্বরূপ, নিম্নলিখিত মাল্টি-ক্লাস ক্লাসিফিকেশন ডেটাসেটটিও শ্রেণী-ভারসাম্যহীন কারণ একটি লেবেলে অন্য দুটির চেয়ে অনেক বেশি উদাহরণ রয়েছে:

  • "সবুজ" শ্রেণী সহ 1,000,000 লেবেল
  • "বেগুনি" ক্লাস সহ 200টি লেবেল
  • "কমলা" ক্লাস সহ 350টি লেবেল

এছাড়াও এনট্রপি , সংখ্যাগরিষ্ঠ শ্রেণী এবং সংখ্যালঘু শ্রেণী দেখুন।

ক্লিপিং

#মৌলিক

নিম্নলিখিত যে কোনো একটি বা উভয়টি করে বহিরাগতদের পরিচালনা করার একটি কৌশল:

  • সর্বাধিক থ্রেশহোল্ডের চেয়ে বেশি বৈশিষ্ট্যের মানগুলিকে সেই সর্বাধিক থ্রেশহোল্ডে হ্রাস করা৷
  • সেই ন্যূনতম থ্রেশহোল্ড পর্যন্ত ন্যূনতম থ্রেশহোল্ডের চেয়ে কম বৈশিষ্ট্যের মানগুলি বৃদ্ধি করা৷

উদাহরণস্বরূপ, ধরুন যে একটি নির্দিষ্ট বৈশিষ্ট্যের <0.5% মান 40-60 রেঞ্জের বাইরে পড়ে। এই ক্ষেত্রে, আপনি নিম্নলিখিত করতে পারেন:

  • 60-এর বেশি (সর্বোচ্চ থ্রেশহোল্ড) সমস্ত মান ঠিক 60 হতে ক্লিপ করুন।
  • 40 এর নিচে সব মান ক্লিপ করুন (সর্বনিম্ন থ্রেশহোল্ড) ঠিক 40 হতে হবে।

বহিরাগতরা মডেলের ক্ষতি করতে পারে, কখনও কখনও প্রশিক্ষণের সময় ওজন উপচে পড়ে। কিছু আউটলিয়ারও নাটকীয়ভাবে নির্ভুলতার মতো মেট্রিক্স নষ্ট করতে পারে। ক্লিপিং ক্ষতি সীমাবদ্ধ করার একটি সাধারণ কৌশল।

গ্রেডিয়েন্ট ক্লিপিং প্রশিক্ষণের সময় একটি নির্দিষ্ট সীমার মধ্যে গ্রেডিয়েন্ট মানগুলিকে জোর করে।

বিভ্রান্তি ম্যাট্রিক্স

#মৌলিক

একটি NxN টেবিল যা একটি শ্রেণীবিভাগ মডেল তৈরি করা সঠিক এবং ভুল ভবিষ্যদ্বাণীগুলির সংখ্যা সংক্ষিপ্ত করে। উদাহরণস্বরূপ, একটি বাইনারি শ্রেণীবিভাগ মডেলের জন্য নিম্নলিখিত বিভ্রান্তি ম্যাট্রিক্স বিবেচনা করুন:

টিউমার (ভবিষ্যদ্বাণী করা) অ-টিউমার (ভবিষ্যদ্বাণী করা)
টিউমার (স্থল সত্য) 18 (TP) 1 (FN)
অ-টিউমার (স্থল সত্য) 6 (FP) 452 (TN)

পূর্ববর্তী বিভ্রান্তি ম্যাট্রিক্স নিম্নলিখিত দেখায়:

  • 19টি ভবিষ্যদ্বাণীর মধ্যে যেখানে গ্রাউন্ড ট্রুথ টিউমার ছিল, মডেলটি সঠিকভাবে 18টি এবং ভুলভাবে 1 শ্রেণীবদ্ধ করেছে।
  • 458টি ভবিষ্যদ্বাণীর মধ্যে যেখানে গ্রাউন্ড ট্রুথ ছিল নন-টিউমার, মডেলটি সঠিকভাবে 452টি এবং ভুলভাবে 6টি শ্রেণীবদ্ধ করেছে।

বহু-শ্রেণীর শ্রেণীবিভাগ সমস্যার জন্য বিভ্রান্তি ম্যাট্রিক্স আপনাকে ভুলের ধরণ সনাক্ত করতে সাহায্য করতে পারে। উদাহরণস্বরূপ, একটি 3-শ্রেণীর বহু-শ্রেণীর শ্রেণীবিভাগ মডেলের জন্য নিম্নলিখিত বিভ্রান্তি ম্যাট্রিক্স বিবেচনা করুন যা তিনটি ভিন্ন আইরিস প্রকার (ভার্জিনিকা, ভার্সিকলার এবং সেটোসা) শ্রেণীবদ্ধ করে। যখন গ্রাউন্ড ট্রুথ ভার্জিনিকা ছিল, তখন কনফিউশন ম্যাট্রিক্স দেখায় যে সেটোসার তুলনায় মডেলটির ভুলভাবে ভার্সিকলার ভবিষ্যদ্বাণী করার সম্ভাবনা অনেক বেশি ছিল:

সেতোসা (ভবিষ্যদ্বাণী করা) ভার্সিকলার (ভবিষ্যদ্বাণী করা) ভার্জিনিকা (ভবিষ্যদ্বাণী করা)
সেতোসা (ভূমি সত্য) ৮৮ 12 0
ভার্সিকলার (ভূমি সত্য) 6 141 7
ভার্জিনিকা (স্থল সত্য) 2 27 109

আরেকটি উদাহরণ হিসাবে, একটি বিভ্রান্তি ম্যাট্রিক্স প্রকাশ করতে পারে যে হাতে লেখা অঙ্কগুলি চিনতে প্রশিক্ষিত একটি মডেল ভুলভাবে 4 এর পরিবর্তে 9 বা ভুলভাবে 7 এর পরিবর্তে 1 ভবিষ্যদ্বাণী করে।

বিভ্রান্তি ম্যাট্রিক্সে যথার্থতা এবং প্রত্যাহার সহ বিভিন্ন কর্মক্ষমতা মেট্রিক্স গণনা করার জন্য পর্যাপ্ত তথ্য রয়েছে।

ক্রমাগত বৈশিষ্ট্য

#মৌলিক

সম্ভাব্য মানের অসীম পরিসর সহ একটি ভাসমান-বিন্দু বৈশিষ্ট্য , যেমন তাপমাত্রা বা ওজন।

পৃথক বৈশিষ্ট্য সঙ্গে বৈসাদৃশ্য.

অভিন্নতা

#মৌলিক

প্রতিটি পুনরাবৃত্তির সাথে ক্ষতির মানগুলি খুব কম বা একেবারেই না পরিবর্তিত হলে এমন একটি অবস্থায় পৌঁছে যায়। উদাহরণস্বরূপ, নিম্নলিখিত ক্ষতির বক্ররেখাটি প্রায় 700 পুনরাবৃত্তিতে একত্রিত হওয়ার পরামর্শ দেয়:

কার্টেসিয়ান প্লট। এক্স-অক্ষ ক্ষতি। Y-অক্ষ হল প্রশিক্ষণের পুনরাবৃত্তির সংখ্যা। প্রথম কয়েকটি পুনরাবৃত্তির সময় ক্ষতি খুব বেশি হয়, তবে দ্রুত হ্রাস পায়। প্রায় 100টি পুনরাবৃত্তির পরে, ক্ষতি এখনও কমছে তবে আরও ধীরে ধীরে। প্রায় 700 পুনরাবৃত্তির পরে, ক্ষতি সমতল থাকে।

একটি মডেল একত্রিত হয় যখন অতিরিক্ত প্রশিক্ষণ মডেল উন্নত করবে না।

গভীর শিক্ষায় , ক্ষতির মান কখনও কখনও স্থির থাকে বা শেষ পর্যন্ত নামার আগে অনেক পুনরাবৃত্তির জন্য প্রায় তাই থাকে। ধ্রুবক ক্ষতির মানগুলির দীর্ঘ সময়ের মধ্যে, আপনি সাময়িকভাবে অভিসারের একটি মিথ্যা ধারণা পেতে পারেন।

তাড়াতাড়ি থামানোও দেখুন।

ডি

ডেটাফ্রেম

#মৌলিক

মেমরিতে ডেটাসেট উপস্থাপনের জন্য একটি জনপ্রিয় পান্ডা ডেটাটাইপ।

একটি ডেটাফ্রেম একটি টেবিল বা স্প্রেডশীটের সাথে সাদৃশ্যপূর্ণ। একটি ডেটাফ্রেমের প্রতিটি কলামের একটি নাম (একটি শিরোনাম) থাকে এবং প্রতিটি সারি একটি অনন্য সংখ্যা দ্বারা চিহ্নিত করা হয়।

একটি ডেটাফ্রেমের প্রতিটি কলাম একটি 2D অ্যারের মতো গঠন করা হয়, প্রতিটি কলামের নিজস্ব ডেটা টাইপ বরাদ্দ করা যেতে পারে।

অফিসিয়াল pandas.DataFrame রেফারেন্স পৃষ্ঠাও দেখুন।

ডেটা সেট বা ডেটাসেট

#মৌলিক

নিম্নোক্ত বিন্যাসগুলির মধ্যে একটিতে সাধারণত (কিন্তু একচেটিয়াভাবে নয়) সংগঠিত কাঁচা ডেটার একটি সংগ্রহ:

  • একটি স্প্রেডশীট
  • CSV (কমা-বিভক্ত মান) বিন্যাসে একটি ফাইল

গভীর মডেল

#মৌলিক

একটি নিউরাল নেটওয়ার্ক যাতে একাধিক লুকানো স্তর থাকে।

একটি গভীর মডেলকে গভীর নিউরাল নেটওয়ার্কও বলা হয়।

প্রশস্ত মডেলের সাথে বৈসাদৃশ্য।

ঘন বৈশিষ্ট্য

#মৌলিক

একটি বৈশিষ্ট্য যেখানে বেশিরভাগ বা সমস্ত মান অশূন্য, সাধারণত ফ্লোটিং-পয়েন্ট মানগুলির একটি টেনসর ৷ উদাহরণস্বরূপ, নিম্নলিখিত 10-উপাদান টেনসরটি ঘন কারণ এর 9টি মান অশূন্য:

8 3 7 5 2 4 0 4 9 6

স্পার্স বৈশিষ্ট্যের সাথে বৈসাদৃশ্য।

গভীরতা

#মৌলিক

একটি নিউরাল নেটওয়ার্কে নিম্নলিখিতগুলির সমষ্টি:

উদাহরণস্বরূপ, পাঁচটি লুকানো স্তর এবং একটি আউটপুট স্তর সহ একটি নিউরাল নেটওয়ার্কের গভীরতা 6।

লক্ষ্য করুন যে ইনপুট স্তর গভীরতা প্রভাবিত করে না।

বিচ্ছিন্ন বৈশিষ্ট্য

#মৌলিক

সম্ভাব্য মানগুলির একটি সীমিত সেট সহ একটি বৈশিষ্ট্য । উদাহরণস্বরূপ, একটি বৈশিষ্ট্য যার মান শুধুমাত্র প্রাণী , উদ্ভিজ্জ বা খনিজ হতে পারে একটি পৃথক (বা শ্রেণীবদ্ধ) বৈশিষ্ট্য।

ক্রমাগত বৈশিষ্ট্য সঙ্গে বৈসাদৃশ্য.

গতিশীল

#মৌলিক

ঘন ঘন বা ক্রমাগত কিছু করা। গতিশীল এবং অনলাইন শব্দগুলি মেশিন লার্নিং এর সমার্থক শব্দ। নিম্নলিখিতগুলি মেশিন লার্নিংয়ে গতিশীল এবং অনলাইনের সাধারণ ব্যবহার:

  • একটি গতিশীল মডেল (বা অনলাইন মডেল ) হল একটি মডেল যা ঘন ঘন বা ক্রমাগত পুনরায় প্রশিক্ষণ দেওয়া হয়।
  • ডায়নামিক ট্রেনিং (বা অনলাইন ট্রেনিং ) হল ঘন ঘন বা একটানা প্রশিক্ষণের প্রক্রিয়া।
  • গতিশীল অনুমান (বা অনলাইন অনুমান ) হল চাহিদার উপর পূর্বাভাস তৈরি করার প্রক্রিয়া।

গতিশীল মডেল

#মৌলিক

একটি মডেল যা ঘন ঘন (সম্ভবত এমনকি ক্রমাগত) পুনরায় প্রশিক্ষিত হয়। একটি গতিশীল মডেল হল একটি "আজীবন শিক্ষার্থী" যা ক্রমাগত বিকশিত ডেটার সাথে খাপ খায়। একটি গতিশীল মডেল একটি অনলাইন মডেল হিসাবেও পরিচিত।

স্ট্যাটিক মডেলের সাথে বৈসাদৃশ্য।

তাড়াতাড়ি থামানো

#মৌলিক

নিয়মিতকরণের একটি পদ্ধতি যা প্রশিক্ষণের ক্ষতি কমানোর আগে প্রশিক্ষণ শেষ করে। প্রারম্ভিক বন্ধে, আপনি ইচ্ছাকৃতভাবে মডেলের প্রশিক্ষণ বন্ধ করে দেন যখন একটি বৈধতা ডেটাসেটের ক্ষতি বাড়তে থাকে; যে, যখন সাধারণীকরণ কর্মক্ষমতা খারাপ হয়.

এম্বেডিং স্তর

#ভাষা
#মৌলিক

একটি বিশেষ লুকানো স্তর যা একটি উচ্চ-মাত্রিক শ্রেণীগত বৈশিষ্ট্যের উপর প্রশিক্ষণ দেয় যা ধীরে ধীরে একটি নিম্ন মাত্রার এম্বেডিং ভেক্টর শিখতে পারে। একটি এম্বেডিং স্তর একটি নিউরাল নেটওয়ার্ককে শুধুমাত্র উচ্চ-মাত্রিক শ্রেণীগত বৈশিষ্ট্যের উপর প্রশিক্ষণের চেয়ে অনেক বেশি দক্ষতার সাথে প্রশিক্ষণ দিতে সক্ষম করে।

উদাহরণস্বরূপ, পৃথিবী বর্তমানে প্রায় 73,000 গাছের প্রজাতিকে সমর্থন করে। ধরুন গাছের প্রজাতি আপনার মডেলের একটি বৈশিষ্ট্য , তাই আপনার মডেলের ইনপুট স্তরে একটি এক-হট ভেক্টর 73,000 উপাদান রয়েছে। উদাহরণস্বরূপ, সম্ভবত baobab এই মত কিছু প্রতিনিধিত্ব করা হবে:

73,000 উপাদানের একটি অ্যারে। প্রথম 6,232টি উপাদান মান 0 ধারণ করে। পরের উপাদানটি 1 মান ধারণ করে। চূড়ান্ত 66,767টি উপাদানের মান শূন্য থাকে।

একটি 73,000-এলিমেন্ট অ্যারে খুব দীর্ঘ। আপনি যদি মডেলটিতে একটি এম্বেডিং স্তর যোগ না করেন, তাহলে 72,999 শূন্য গুণ করার কারণে প্রশিক্ষণটি খুব সময়সাপেক্ষ হতে চলেছে। সম্ভবত আপনি 12টি মাত্রা সমন্বিত করার জন্য এম্বেডিং স্তরটি বেছে নিন। ফলস্বরূপ, এম্বেডিং স্তরটি ধীরে ধীরে প্রতিটি গাছের প্রজাতির জন্য একটি নতুন এমবেডিং ভেক্টর শিখবে।

কিছু পরিস্থিতিতে, হ্যাশিং একটি এম্বেডিং স্তরের একটি যুক্তিসঙ্গত বিকল্প।

যুগ

#মৌলিক

পুরো প্রশিক্ষণ সেটের উপর একটি সম্পূর্ণ প্রশিক্ষণ পাস যাতে প্রতিটি উদাহরণ একবার প্রক্রিয়া করা হয়েছে।

একটি যুগ N / ব্যাচ আকারের প্রশিক্ষণের পুনরাবৃত্তির প্রতিনিধিত্ব করে, যেখানে N হল মোট উদাহরণের সংখ্যা।

উদাহরণস্বরূপ, নিম্নলিখিতটি ধরুন:

  • ডেটাসেটটিতে 1,000টি উদাহরণ রয়েছে।
  • ব্যাচ আকার 50 উদাহরণ.

অতএব, একটি একক যুগের জন্য 20টি পুনরাবৃত্তি প্রয়োজন:

1 epoch = (N/batch size) = (1,000 / 50) = 20 iterations

উদাহরণ

#মৌলিক

বৈশিষ্ট্যের একটি সারির মান এবং সম্ভবত একটি লেবেলতত্ত্বাবধানে শিক্ষার উদাহরণ দুটি সাধারণ বিভাগে পড়ে:

  • একটি লেবেলযুক্ত উদাহরণে এক বা একাধিক বৈশিষ্ট্য এবং একটি লেবেল থাকে। লেবেলযুক্ত উদাহরণ প্রশিক্ষণের সময় ব্যবহার করা হয়।
  • লেবেলবিহীন উদাহরণে এক বা একাধিক বৈশিষ্ট্য থাকে কিন্তু কোনো লেবেল থাকে না। লেবেলবিহীন উদাহরণ অনুমানের সময় ব্যবহার করা হয়।

উদাহরণস্বরূপ, ধরুন আপনি শিক্ষার্থীদের পরীক্ষার স্কোরের উপর আবহাওয়ার অবস্থার প্রভাব নির্ধারণের জন্য একটি মডেলকে প্রশিক্ষণ দিচ্ছেন। এখানে তিনটি লেবেলযুক্ত উদাহরণ রয়েছে:

বৈশিষ্ট্য লেবেল
তাপমাত্রা আর্দ্রতা চাপ পরীক্ষার স্কোর
15 47 998 ভাল
19 34 1020 চমৎকার
18 92 1012 দরিদ্র

এখানে তিনটি লেবেলবিহীন উদাহরণ রয়েছে:

তাপমাত্রা আর্দ্রতা চাপ
12 62 1014
21 47 1017
19 41 1021

একটি ডেটাসেটের সারিটি সাধারণত একটি উদাহরণের জন্য কাঁচা উত্স। অর্থাৎ, একটি উদাহরণে সাধারণত ডেটাসেটের কলামগুলির একটি উপসেট থাকে। উপরন্তু, একটি উদাহরণের বৈশিষ্ট্যগুলিতে সিন্থেটিক বৈশিষ্ট্যগুলিও অন্তর্ভুক্ত থাকতে পারে, যেমন বৈশিষ্ট্য ক্রস

মিথ্যা নেতিবাচক (এফএন)

#মৌলিক

একটি উদাহরণ যেখানে মডেলটি ভুলভাবে নেতিবাচক শ্রেণীর ভবিষ্যদ্বাণী করে। উদাহরণস্বরূপ, মডেলটি ভবিষ্যদ্বাণী করে যে একটি নির্দিষ্ট ইমেল বার্তা স্প্যাম নয় (নেতিবাচক শ্রেণি), কিন্তু সেই ইমেল বার্তাটি আসলে স্প্যাম

মিথ্যা পজিটিভ (FP)

#মৌলিক

একটি উদাহরণ যেখানে মডেল ভুলভাবে ইতিবাচক শ্রেণীর ভবিষ্যদ্বাণী করে। উদাহরণস্বরূপ, মডেলটি ভবিষ্যদ্বাণী করে যে একটি নির্দিষ্ট ইমেল বার্তাটি স্প্যাম (পজিটিভ ক্লাস), কিন্তু সেই ইমেল বার্তাটি আসলে স্প্যাম নয়

মিথ্যা ইতিবাচক হার (FPR)

#মৌলিক

প্রকৃত নেতিবাচক উদাহরণের অনুপাত যার জন্য মডেলটি ভুলভাবে ইতিবাচক শ্রেণীর ভবিষ্যদ্বাণী করেছে। নিম্নলিখিত সূত্রটি মিথ্যা ইতিবাচক হার গণনা করে:

$$\text{false positive rate} = \frac{\text{false positives}}{\text{false positives} + \text{true negatives}}$$

মিথ্যা ধনাত্মক হার হল একটি ROC বক্ররেখার x-অক্ষ।

বৈশিষ্ট্য

#মৌলিক

একটি মেশিন লার্নিং মডেলের একটি ইনপুট পরিবর্তনশীল। একটি উদাহরণ এক বা একাধিক বৈশিষ্ট্য নিয়ে গঠিত। উদাহরণস্বরূপ, ধরুন আপনি শিক্ষার্থীদের পরীক্ষার স্কোরের উপর আবহাওয়ার অবস্থার প্রভাব নির্ধারণের জন্য একটি মডেলকে প্রশিক্ষণ দিচ্ছেন। নিম্নলিখিত সারণী তিনটি উদাহরণ দেখায়, যার প্রতিটিতে তিনটি বৈশিষ্ট্য এবং একটি লেবেল রয়েছে:

বৈশিষ্ট্য লেবেল
তাপমাত্রা আর্দ্রতা চাপ পরীক্ষার স্কোর
15 47 998 92
19 34 1020 84
18 92 1012 87

লেবেলের সাথে বৈসাদৃশ্য।

বৈশিষ্ট্য ক্রস

#মৌলিক

একটি সিন্থেটিক বৈশিষ্ট্য "ক্রসিং" শ্রেণীবদ্ধ বা বালতি বৈশিষ্ট্য দ্বারা গঠিত।

উদাহরণস্বরূপ, একটি "মেজাজ পূর্বাভাস" মডেল বিবেচনা করুন যা নিম্নলিখিত চারটি বালতিগুলির মধ্যে একটিতে তাপমাত্রা উপস্থাপন করে:

  • freezing
  • chilly
  • temperate
  • warm

এবং নিম্নলিখিত তিনটি বালতিগুলির মধ্যে একটিতে বাতাসের গতি উপস্থাপন করে:

  • still
  • light
  • windy

বৈশিষ্ট্য ক্রস ছাড়াই, রৈখিক মডেল পূর্ববর্তী সাতটি বিভিন্ন বালতিগুলির প্রতিটিতে স্বাধীনভাবে ট্রেন করে। সুতরাং, মডেলটি ট্রেনিং করে, উদাহরণস্বরূপ, প্রশিক্ষণের থেকে স্বাধীনভাবে freezing , উদাহরণস্বরূপ, windy

বিকল্পভাবে, আপনি তাপমাত্রা এবং বাতাসের গতির একটি বৈশিষ্ট্য ক্রস তৈরি করতে পারেন। এই সিন্থেটিক বৈশিষ্ট্যের নিম্নলিখিত 12টি সম্ভাব্য মান থাকবে:

  • freezing-still
  • freezing-light
  • freezing-windy
  • chilly-still
  • chilly-light
  • chilly-windy
  • temperate-still
  • temperate-light
  • temperate-windy
  • warm-still
  • warm-light
  • warm-windy

ফিচার ক্রসের জন্য ধন্যবাদ, মডেলটি freezing-windy দিন এবং freezing-still দিনের মধ্যে মেজাজের পার্থক্য শিখতে পারে।

আপনি যদি দুটি বৈশিষ্ট্য থেকে একটি সিন্থেটিক বৈশিষ্ট্য তৈরি করেন যার প্রতিটিতে অনেকগুলি আলাদা বালতি থাকে, ফলে বৈশিষ্ট্য ক্রসটিতে বিপুল সংখ্যক সম্ভাব্য সংমিশ্রণ থাকবে। উদাহরণস্বরূপ, যদি একটি বৈশিষ্ট্যে 1,000 বালতি থাকে এবং অন্য বৈশিষ্ট্যটিতে 2,000টি বালতি থাকে, ফলে বৈশিষ্ট্য ক্রসটিতে 2,000,000 বালতি থাকে৷

আনুষ্ঠানিকভাবে, একটি ক্রস একটি কার্টেসিয়ান পণ্য

ফিচার ক্রসগুলি বেশিরভাগ লিনিয়ার মডেলের সাথে ব্যবহৃত হয় এবং খুব কমই নিউরাল নেটওয়ার্কের সাথে ব্যবহার করা হয়।

বৈশিষ্ট্য প্রকৌশল

#মৌলিক
#টেনসরফ্লো

একটি প্রক্রিয়া যা নিম্নলিখিত পদক্ষেপগুলিকে অন্তর্ভুক্ত করে:

  1. মডেল প্রশিক্ষণের জন্য কোন বৈশিষ্ট্যগুলি কার্যকর হতে পারে তা নির্ধারণ করা।
  2. ডেটাসেট থেকে কাঁচা ডেটাকে সেই বৈশিষ্ট্যগুলির দক্ষ সংস্করণে রূপান্তর করা।

উদাহরণস্বরূপ, আপনি নির্ধারণ করতে পারেন যে temperature একটি দরকারী বৈশিষ্ট্য হতে পারে। তারপর, মডেলটি বিভিন্ন temperature রেঞ্জ থেকে কী শিখতে পারে তা অপ্টিমাইজ করতে আপনি বাকেটিংয়ের সাথে পরীক্ষা করতে পারেন।

ফিচার ইঞ্জিনিয়ারিংকে কখনও কখনও ফিচার এক্সট্রাকশন বলা হয়।

বৈশিষ্ট্য সেট

#মৌলিক

আপনার মেশিন লার্নিং মডেল ট্রেনের বৈশিষ্ট্যগুলির গ্রুপ। উদাহরণস্বরূপ, পোস্টাল কোড, সম্পত্তির আকার এবং সম্পত্তির অবস্থার মধ্যে একটি মডেলের জন্য একটি সাধারণ বৈশিষ্ট্য সেট থাকতে পারে যা আবাসন মূল্যের পূর্বাভাস দেয়।

বৈশিষ্ট্য ভেক্টর

#মৌলিক

একটি উদাহরণ সমন্বিত বৈশিষ্ট্য মানের অ্যারে। প্রশিক্ষণের সময় এবং অনুমানের সময় বৈশিষ্ট্য ভেক্টর ইনপুট হয়। উদাহরণস্বরূপ, দুটি পৃথক বৈশিষ্ট্য সহ একটি মডেলের বৈশিষ্ট্য ভেক্টর হতে পারে:

[0.92, 0.56]

চারটি স্তর: একটি ইনপুট স্তর, দুটি লুকানো স্তর এবং একটি আউটপুট স্তর। ইনপুট স্তরে দুটি নোড রয়েছে, একটিতে 0.92 মান এবং অন্যটিতে 0.56 মান রয়েছে।

প্রতিটি উদাহরণ বৈশিষ্ট্য ভেক্টরের জন্য বিভিন্ন মান সরবরাহ করে, তাই পরবর্তী উদাহরণের জন্য বৈশিষ্ট্য ভেক্টর এমন কিছু হতে পারে:

[0.73, 0.49]

ফিচার ইঞ্জিনিয়ারিং ফিচার ভেক্টরে বৈশিষ্ট্যগুলিকে কীভাবে উপস্থাপন করতে হয় তা নির্ধারণ করে। উদাহরণস্বরূপ, পাঁচটি সম্ভাব্য মান সহ একটি বাইনারি শ্রেণীবদ্ধ বৈশিষ্ট্য এক-হট এনকোডিং দিয়ে উপস্থাপন করা যেতে পারে। এই ক্ষেত্রে, একটি নির্দিষ্ট উদাহরণের জন্য বৈশিষ্ট্য ভেক্টরের অংশে চারটি শূন্য এবং তৃতীয় অবস্থানে একটি একক 1.0 থাকবে, নিম্নরূপ:

[0.0, 0.0, 1.0, 0.0, 0.0]

অন্য উদাহরণ হিসাবে, ধরুন আপনার মডেল তিনটি বৈশিষ্ট্য নিয়ে গঠিত:

  • এক-হট এনকোডিং দ্বারা উপস্থাপিত পাঁচটি সম্ভাব্য মান সহ একটি বাইনারি শ্রেণীবদ্ধ বৈশিষ্ট্য; for example: [0.0, 1.0, 0.0, 0.0, 0.0]
  • another binary categorical feature with three possible values represented with one-hot encoding; for example: [0.0, 0.0, 1.0]
  • a floating-point feature; for example: 8.3 .

In this case, the feature vector for each example would be represented by nine values. Given the example values in the preceding list, the feature vector would be:

0.0
1.0
0.0
0.0
0.0
0.0
0.0
1.0
8.3

feedback loop

#fundamentals

In machine learning, a situation in which a model's predictions influence the training data for the same model or another model. For example, a model that recommends movies will influence the movies that people see, which will then influence subsequent movie recommendation models.

জি

সাধারণীকরণ

#fundamentals

A model's ability to make correct predictions on new, previously unseen data. A model that can generalize is the opposite of a model that is overfitting .

generalization curve

#fundamentals

A plot of both training loss and validation loss as a function of the number of iterations .

A generalization curve can help you detect possible overfitting . For example, the following generalization curve suggests overfitting because validation loss ultimately becomes significantly higher than training loss.

A Cartesian graph in which the y-axis is labeled 'loss' and the x-axis
          is labeled 'iterations'. Two plots appear. One plots shows the
          training loss and the other shows the validation loss.
          The two plots start off similarly, but the training loss eventually
          dips far lower than the validation loss.

গ্রেডিয়েন্ট ডিসেন্ট

#fundamentals

A mathematical technique to minimize loss . Gradient descent iteratively adjusts weights and biases , gradually finding the best combination to minimize loss.

Gradient descent is older—much, much older—than machine learning.

কঠিন সত্য

#fundamentals

বাস্তবতা।

The thing that actually happened.

For example, consider a binary classification model that predicts whether a student in their first year of university will graduate within six years. Ground truth for this model is whether or not that student actually graduated within six years.

এইচ

লুকানো স্তর

#fundamentals

A layer in a neural network between the input layer (the features) and the output layer (the prediction). Each hidden layer consists of one or more neurons . For example, the following neural network contains two hidden layers, the first with three neurons and the second with two neurons:

Four layers. The first layer is an input layer containing two
          features. The second layer is a hidden layer containing three
          neurons. The third layer is a hidden layer containing two
          neurons. The fourth layer is an output layer. Each feature
          contains three edges, each of which points to a different neuron
          in the second layer. Each of the neurons in the second layer
          contains two edges, each of which points to a different neuron
          in the third layer. Each of the neurons in the third layer contain
          one edge, each pointing to the output layer.

A deep neural network contains more than one hidden layer. For example, the preceding illustration is a deep neural network because the model contains two hidden layers.

hyperparameter

#fundamentals

The variables that you or a hyperparameter tuning serviceadjust during successive runs of training a model. For example, learning rate is a hyperparameter. You could set the learning rate to 0.01 before one training session. If you determine that 0.01 is too high, you could perhaps set the learning rate to 0.003 for the next training session.

In contrast, parameters are the various weights and bias that the model learns during training.

আমি

independently and identically distributed (iid)

#fundamentals

Data drawn from a distribution that doesn't change, and where each value drawn doesn't depend on values that have been drawn previously. An iid is the ideal gas of machine learning—a useful mathematical construct but almost never exactly found in the real world. For example, the distribution of visitors to a web page may be iid over a brief window of time; that is, the distribution doesn't change during that brief window and one person's visit is generally independent of another's visit. However, if you expand that window of time, seasonal differences in the web page's visitors may appear.

See also nonstationarity .

অনুমান

#fundamentals

In machine learning, the process of making predictions by applying a trained model to unlabeled examples .

Inference has a somewhat different meaning in statistics. See the Wikipedia article on statistical inference for details.

input layer

#fundamentals

The layer of a neural network that holds the feature vector . That is, the input layer provides examples for training or inference . For example, the input layer in the following neural network consists of two features:

Four layers: an input layer, two hidden layers, and an output layer.

interpretability

#fundamentals

The ability to explain or to present an ML model's reasoning in understandable terms to a human.

Most linear regression models, for example, are highly interpretable. (You merely need to look at the trained weights for each feature.) Decision forests are also highly interpretable. Some models, however, require sophisticated visualization to become interpretable.

You can use the Learning Interpretability Tool (LIT) to interpret ML models.

পুনরাবৃত্তি

#fundamentals

A single update of a model's parameters—the model's weights and biases —during training . The batch size determines how many examples the model processes in a single iteration. For instance, if the batch size is 20, then the model processes 20 examples before adjusting the parameters.

When training a neural network , a single iteration involves the following two passes:

  1. A forward pass to evaluate loss on a single batch.
  2. A backward pass ( backpropagation ) to adjust the model's parameters based on the loss and the learning rate.

এল

L 0 regularization

#fundamentals

A type of regularization that penalizes the total number of nonzero weights in a model. For example, a model having 11 nonzero weights would be penalized more than a similar model having 10 nonzero weights.

L 0 regularization is sometimes called L0-norm regularization .

L 1 loss

#fundamentals

A loss function that calculates the absolute value of the difference between actual label values and the values that a model predicts. For example, here's the calculation of L 1 loss for a batch of five examples :

Actual value of example Model's predicted value Absolute value of delta
7 6 1
5 4 1
8 11 3
4 6 2
9 8 1
8 = L 1 loss

L 1 loss is less sensitive to outliers than L 2 loss .

The Mean Absolute Error is the average L 1 loss per example.

L 1 regularization

#fundamentals

A type of regularization that penalizes weights in proportion to the sum of the absolute value of the weights. L 1 regularization helps drive the weights of irrelevant or barely relevant features to exactly 0 . A feature with a weight of 0 is effectively removed from the model.

Contrast with L 2 regularization .

L 2 loss

#fundamentals

A loss function that calculates the square of the difference between actual label values and the values that a model predicts. For example, here's the calculation of L 2 loss for a batch of five examples :

Actual value of example Model's predicted value Square of delta
7 6 1
5 4 1
8 11 9
4 6 4
9 8 1
16 = L 2 loss

Due to squaring, L 2 loss amplifies the influence of outliers . That is, L 2 loss reacts more strongly to bad predictions than L 1 loss . For example, the L 1 loss for the preceding batch would be 8 rather than 16. Notice that a single outlier accounts for 9 of the 16.

Regression models typically use L 2 loss as the loss function.

The Mean Squared Error is the average L 2 loss per example. Squared loss is another name for L 2 loss.

L 2 regularization

#fundamentals

A type of regularization that penalizes weights in proportion to the sum of the squares of the weights. L 2 regularization helps drive outlier weights (those with high positive or low negative values) closer to 0 but not quite to 0 . Features with values very close to 0 remain in the model but don't influence the model's prediction very much.

L 2 regularization always improves generalization in linear models .

Contrast with L 1 regularization .

লেবেল

#fundamentals

In supervised machine learning , the "answer" or "result" portion of an example .

Each labeled example consists of one or more features and a label. For instance, in a spam detection dataset, the label would probably be either "spam" or "not spam." In a rainfall dataset, the label might be the amount of rain that fell during a certain period.

labeled example

#fundamentals

An example that contains one or more features and a label . For example, the following table shows three labeled examples from a house valuation model, each with three features and one label:

বেডরুমের সংখ্যা বাথরুমের সংখ্যা House age House price (label)
3 2 15 $345,000
2 1 72 $179,000
4 2 34 $392,000

In supervised machine learning , models train on labeled examples and make predictions on unlabeled examples .

Contrast labeled example with unlabeled examples.

ল্যাম্বডা

#fundamentals

Synonym for regularization rate .

Lambda is an overloaded term. Here we're focusing on the term's definition within regularization .

স্তর

#fundamentals

A set of neurons in a neural network . Three common types of layers are as follows:

For example, the following illustration shows a neural network with one input layer, two hidden layers, and one output layer:

A neural network with one input layer, two hidden layers, and one
          output layer. The input layer consists of two features. The first
          hidden layer consists of three neurons and the second hidden layer
          consists of two neurons. The output layer consists of a single node.

In TensorFlow , layers are also Python functions that take Tensors and configuration options as input and produce other tensors as output.

শেখার হার

#fundamentals

A floating-point number that tells the gradient descent algorithm how strongly to adjust weights and biases on each iteration . For example, a learning rate of 0.3 would adjust weights and biases three times more powerfully than a learning rate of 0.1.

Learning rate is a key hyperparameter . If you set the learning rate too low, training will take too long. If you set the learning rate too high, gradient descent often has trouble reaching convergence .

রৈখিক

#fundamentals

A relationship between two or more variables that can be represented solely through addition and multiplication.

The plot of a linear relationship is a line.

Contrast with nonlinear .

linear model

#fundamentals

A model that assigns one weight per feature to make predictions . (Linear models also incorporate a bias .) In contrast, the relationship of features to predictions in deep models is generally nonlinear .

Linear models are usually easier to train and more interpretable than deep models. However, deep models can learn complex relationships between features.

Linear regression and logistic regression are two types of linear models.

লিনিয়ার রিগ্রেশন

#fundamentals

A type of machine learning model in which both of the following are true:

  • The model is a linear model .
  • The prediction is a floating-point value. (This is the regression part of linear regression .)

Contrast linear regression with logistic regression . Also, contrast regression with classification .

logistic regression

#fundamentals

A type of regression model that predicts a probability. Logistic regression models have the following characteristics:

  • The label is categorical . The term logistic regression usually refers to binary logistic regression , that is, to a model that calculates probabilities for labels with two possible values. A less common variant, multinomial logistic regression , calculates probabilities for labels with more than two possible values.
  • The loss function during training is Log Loss . (Multiple Log Loss units can be placed in parallel for labels with more than two possible values.)
  • The model has a linear architecture, not a deep neural network. However, the remainder of this definition also applies to deep models that predict probabilities for categorical labels.

For example, consider a logistic regression model that calculates the probability of an input email being either spam or not spam. During inference, suppose the model predicts 0.72. Therefore, the model is estimating:

  • A 72% chance of the email being spam.
  • A 28% chance of the email not being spam.

A logistic regression model uses the following two-step architecture:

  1. The model generates a raw prediction (y') by applying a linear function of input features.
  2. The model uses that raw prediction as input to a sigmoid function , which converts the raw prediction to a value between 0 and 1, exclusive.

Like any regression model, a logistic regression model predicts a number. However, this number typically becomes part of a binary classification model as follows:

  • If the predicted number is greater than the classification threshold , the binary classification model predicts the positive class.
  • If the predicted number is less than the classification threshold, the binary classification model predicts the negative class.

লগ ক্ষতি

#fundamentals

The loss function used in binary logistic regression .

log-odds

#fundamentals

The logarithm of the odds of some event.

ক্ষতি

#fundamentals

During the training of a supervised model , a measure of how far a model's prediction is from its label .

A loss function calculates the loss.

loss curve

#fundamentals

A plot of loss as a function of the number of training iterations . The following plot shows a typical loss curve:

A Cartesian graph of loss versus training iterations, showing a
          rapid drop in loss for the initial iterations, followed by a gradual
          drop, and then a flat slope during the final iterations.

Loss curves can help you determine when your model is converging or overfitting .

Loss curves can plot all of the following types of loss:

See also generalization curve .

ক্ষতি ফাংশন

#fundamentals

During training or testing, a mathematical function that calculates the loss on a batch of examples. A loss function returns a lower loss for models that makes good predictions than for models that make bad predictions.

The goal of training is typically to minimize the loss that a loss function returns.

Many different kinds of loss functions exist. Pick the appropriate loss function for the kind of model you are building. উদাহরণ স্বরূপ:

এম

মেশিন লার্নিং

#fundamentals

A program or system that trains a model from input data. The trained model can make useful predictions from new (never-before-seen) data drawn from the same distribution as the one used to train the model.

Machine learning also refers to the field of study concerned with these programs or systems.

majority class

#fundamentals

The more common label in a class-imbalanced dataset . For example, given a dataset containing 99% negative labels and 1% positive labels, the negative labels are the majority class.

Contrast with minority class .

mini-batch

#fundamentals

A small, randomly selected subset of a batch processed in one iteration . The batch size of a mini-batch is usually between 10 and 1,000 examples.

For example, suppose the entire training set (the full batch) consists of 1,000 examples. Further suppose that you set the batch size of each mini-batch to 20. Therefore, each iteration determines the loss on a random 20 of the 1,000 examples and then adjusts the weights and biases accordingly.

It is much more efficient to calculate the loss on a mini-batch than the loss on all the examples in the full batch.

minority class

#fundamentals

The less common label in a class-imbalanced dataset . For example, given a dataset containing 99% negative labels and 1% positive labels, the positive labels are the minority class.

Contrast with majority class .

মডেল

#fundamentals

In general, any mathematical construct that processes input data and returns output. Phrased differently, a model is the set of parameters and structure needed for a system to make predictions. In supervised machine learning , a model takes an example as input and infers a prediction as output. Within supervised machine learning, models differ somewhat. উদাহরণ স্বরূপ:

  • A linear regression model consists of a set of weights and a bias .
  • A neural network model consists of:
    • A set of hidden layers , each containing one or more neurons .
    • The weights and bias associated with each neuron.
  • A decision tree model consists of:
    • The shape of the tree; that is, the pattern in which the conditions and leaves are connected.
    • The conditions and leaves.

You can save, restore, or make copies of a model.

Unsupervised machine learning also generates models, typically a function that can map an input example to the most appropriate cluster .

multi-class classification

#fundamentals

In supervised learning, a classification problem in which the dataset contains more than two classes of labels. For example, the labels in the Iris dataset must be one of the following three classes:

  • Iris setosa
  • Iris virginica
  • Iris versicolor

A model trained on the Iris dataset that predicts Iris type on new examples is performing multi-class classification.

In contrast, classification problems that distinguish between exactly two classes are binary classification models . For example, an email model that predicts either spam or not spam is a binary classification model.

In clustering problems, multi-class classification refers to more than two clusters.

এন

negative class

#fundamentals

In binary classification , one class is termed positive and the other is termed negative . The positive class is the thing or event that the model is testing for and the negative class is the other possibility. উদাহরণ স্বরূপ:

  • The negative class in a medical test might be "not tumor."
  • The negative class in an email classifier might be "not spam."

Contrast with positive class .

নিউরাল নেটওয়ার্ক

#fundamentals

A model containing at least one hidden layer . A deep neural network is a type of neural network containing more than one hidden layer. For example, the following diagram shows a deep neural network containing two hidden layers.

A neural network with an input layer, two hidden layers, and an
          output layer.

Each neuron in a neural network connects to all of the nodes in the next layer. For example, in the preceding diagram, notice that each of the three neurons in the first hidden layer separately connect to both of the two neurons in the second hidden layer.

Neural networks implemented on computers are sometimes called artificial neural networks to differentiate them from neural networks found in brains and other nervous systems.

Some neural networks can mimic extremely complex nonlinear relationships between different features and the label.

See also convolutional neural network and recurrent neural network .

নিউরন

#fundamentals

In machine learning, a distinct unit within a hidden layer of a neural network . Each neuron performs the following two-step action:

  1. Calculates the weighted sum of input values multiplied by their corresponding weights.
  2. Passes the weighted sum as input to an activation function .

A neuron in the first hidden layer accepts inputs from the feature values in the input layer . A neuron in any hidden layer beyond the first accepts inputs from the neurons in the preceding hidden layer. For example, a neuron in the second hidden layer accepts inputs from the neurons in the first hidden layer.

The following illustration highlights two neurons and their inputs.

A neural network with an input layer, two hidden layers, and an
          output layer. Two neurons are highlighted: one in the first
          hidden layer and one in the second hidden layer. The highlighted
          neuron in the first hidden layer receives inputs from both features
          in the input layer. The highlighted neuron in the second hidden layer
          receives inputs from each of the three neurons in the first hidden
          layer.

A neuron in a neural network mimics the behavior of neurons in brains and other parts of nervous systems.

node (neural network)

#fundamentals

A neuron in a hidden layer .

অরৈখিক

#fundamentals

A relationship between two or more variables that can't be represented solely through addition and multiplication. A linear relationship can be represented as a line; a nonlinear relationship can't be represented as a line. For example, consider two models that each relate a single feature to a single label. The model on the left is linear and the model on the right is nonlinear:

Two plots. One plot is a line, so this is a linear relationship.
          The other plot is a curve, so this is a nonlinear relationship.

nonstationarity

#fundamentals

A feature whose values change across one or more dimensions, usually time. For example, consider the following examples of nonstationarity:

  • The number of swimsuits sold at a particular store varies with the season.
  • The quantity of a particular fruit harvested in a particular region is zero for much of the year but large for a brief period.
  • Due to climate change, annual mean temperatures are shifting.

Contrast with stationarity .

স্বাভাবিকীকরণ

#fundamentals

Broadly speaking, the process of converting a variable's actual range of values into a standard range of values, such as:

  • -1 to +1
  • 0 থেকে 1
  • the normal distribution

For example, suppose the actual range of values of a certain feature is 800 to 2,400. As part of feature engineering , you could normalize the actual values down to a standard range, such as -1 to +1.

Normalization is a common task in feature engineering . Models usually train faster (and produce better predictions) when every numerical feature in the feature vector has roughly the same range.

গানিতিক তথ্য

#fundamentals

Features represented as integers or real-valued numbers. For example, a house valuation model would probably represent the size of a house (in square feet or square meters) as numerical data. Representing a feature as numerical data indicates that the feature's values have a mathematical relationship to the label. That is, the number of square meters in a house probably has some mathematical relationship to the value of the house.

Not all integer data should be represented as numerical data. For example, postal codes in some parts of the world are integers; however, integer postal codes should not be represented as numerical data in models. That's because a postal code of 20000 is not twice (or half) as potent as a postal code of 10000. Furthermore, although different postal codes do correlate to different real estate values, we can't assume that real estate values at postal code 20000 are twice as valuable as real estate values at postal code 10000. Postal codes should be represented as categorical data instead.

Numerical features are sometimes called continuous features .

অফলাইন

#fundamentals

Synonym for static .

offline inference

#fundamentals

The process of a model generating a batch of predictions and then caching (saving) those predictions. Apps can then access the desired prediction from the cache rather than rerunning the model.

For example, consider a model that generates local weather forecasts (predictions) once every four hours. After each model run, the system caches all the local weather forecasts. Weather apps retrieve the forecasts from the cache.

Offline inference is also called static inference .

Contrast with online inference .

one-hot encoding

#fundamentals

Representing categorical data as a vector in which:

  • One element is set to 1.
  • All other elements are set to 0.

One-hot encoding is commonly used to represent strings or identifiers that have a finite set of possible values. For example, suppose a certain categorical feature named Scandinavia has five possible values:

  • "Denmark"
  • "সুইডেন"
  • "নরওয়ে"
  • "ফিনল্যান্ড"
  • "আইসল্যান্ড"

One-hot encoding could represent each of the five values as follows:

দেশ ভেক্টর
"Denmark" 1 0 0 0 0
"সুইডেন" 0 1 0 0 0
"নরওয়ে" 0 0 1 0 0
"ফিনল্যান্ড" 0 0 0 1 0
"আইসল্যান্ড" 0 0 0 0 1

Thanks to one-hot encoding, a model can learn different connections based on each of the five countries.

Representing a feature as numerical data is an alternative to one-hot encoding. Unfortunately, representing the Scandinavian countries numerically is not a good choice. For example, consider the following numeric representation:

  • "Denmark" is 0
  • "Sweden" is 1
  • "Norway" is 2
  • "Finland" is 3
  • "Iceland" is 4

With numeric encoding, a model would interpret the raw numbers mathematically and would try to train on those numbers. However, Iceland isn't actually twice as much (or half as much) of something as Norway, so the model would come to some strange conclusions.

one-vs.-all

#fundamentals

Given a classification problem with N classes, a solution consisting of N separate binary classifiers —one binary classifier for each possible outcome. For example, given a model that classifies examples as animal, vegetable, or mineral, a one-vs.-all solution would provide the following three separate binary classifiers:

  • animal vs. not animal
  • vegetable vs. not vegetable
  • mineral vs. not mineral

অনলাইন

#fundamentals

Synonym for dynamic .

online inference

#fundamentals

Generating predictions on demand. For example, suppose an app passes input to a model and issues a request for a prediction. A system using online inference responds to the request by running the model (and returning the prediction to the app).

Contrast with offline inference .

output layer

#fundamentals

The "final" layer of a neural network. The output layer contains the prediction.

The following illustration shows a small deep neural network with an input layer, two hidden layers, and an output layer:

A neural network with one input layer, two hidden layers, and one
          output layer. The input layer consists of two features. The first
          hidden layer consists of three neurons and the second hidden layer
          consists of two neurons. The output layer consists of a single node.

ওভারফিটিং

#fundamentals

Creating a model that matches the training data so closely that the model fails to make correct predictions on new data.

Regularization can reduce overfitting. Training on a large and diverse training set can also reduce overfitting.

পৃ

পান্ডা

#fundamentals

A column-oriented data analysis API built on top of numpy . Many machine learning frameworks, including TensorFlow, support pandas data structures as inputs. See the pandas documentation for details.

প্যারামিটার

#fundamentals

The weights and biases that a model learns during training . For example, in a linear regression model, the parameters consist of the bias ( b ) and all the weights ( w 1 , w 2 , and so on) in the following formula:

$$y' = b + w_1x_1 + w_2x_2 + … w_nx_n$$

In contrast, hyperparameter are the values that you (or a hyperparameter turning service) supply to the model. For example, learning rate is a hyperparameter.

positive class

#fundamentals

The class you are testing for.

For example, the positive class in a cancer model might be "tumor." The positive class in an email classifier might be "spam."

Contrast with negative class .

পোস্ট প্রসেসিং

#fairness
#fundamentals

Adjusting the output of a model after the model has been run. Post-processing can be used to enforce fairness constraints without modifying models themselves.

For example, one might apply post-processing to a binary classifier by setting a classification threshold such that equality of opportunity is maintained for some attribute by checking that the true positive rate is the same for all values of that attribute.

ভবিষ্যদ্বাণী

#fundamentals

A model's output. উদাহরণ স্বরূপ:

  • The prediction of a binary classification model is either the positive class or the negative class.
  • The prediction of a multi-class classification model is one class.
  • The prediction of a linear regression model is a number.

proxy labels

#fundamentals

Data used to approximate labels not directly available in a dataset.

For example, suppose you must train a model to predict employee stress level. Your dataset contains a lot of predictive features but doesn't contain a label named stress level. Undaunted, you pick "workplace accidents" as a proxy label for stress level. After all, employees under high stress get into more accidents than calm employees. নাকি তারা করে? Maybe workplace accidents actually rise and fall for multiple reasons.

As a second example, suppose you want is it raining? to be a Boolean label for your dataset, but your dataset doesn't contain rain data. If photographs are available, you might establish pictures of people carrying umbrellas as a proxy label for is it raining? Is that a good proxy label? Possibly, but people in some cultures may be more likely to carry umbrellas to protect against sun than the rain.

Proxy labels are often imperfect. When possible, choose actual labels over proxy labels. That said, when an actual label is absent, pick the proxy label very carefully, choosing the least horrible proxy label candidate.

আর

রেটার

#fundamentals

A human who provides labels for examples . "Annotator" is another name for rater.

সংশোধনকৃত লিনিয়ার ইউনিট (ReLU)

#fundamentals

An activation function with the following behavior:

  • If input is negative or zero, then the output is 0.
  • If input is positive, then the output is equal to the input.

উদাহরণ স্বরূপ:

  • If the input is -3, then the output is 0.
  • If the input is +3, then the output is 3.0.

Here is a plot of ReLU:

A cartesian plot of two lines. The first line has a constant
          y value of 0, running along the x-axis from -infinity,0 to 0,-0.
          The second line starts at 0,0. This line has a slope of +1, so
          it runs from 0,0 to +infinity,+infinity.

ReLU is a very popular activation function. Despite its simple behavior, ReLU still enables a neural network to learn nonlinear relationships between features and the label .

regression model

#fundamentals

Informally, a model that generates a numerical prediction. (In contrast, a classification model generates a class prediction.) For example, the following are all regression models:

  • A model that predicts a certain house's value, such as 423,000 Euros.
  • A model that predicts a certain tree's life expectancy, such as 23.2 years.
  • A model that predicts the amount of rain that will fall in a certain city over the next six hours, such as 0.18 inches.

Two common types of regression models are:

  • Linear regression , which finds the line that best fits label values to features.
  • Logistic regression , which generates a probability between 0.0 and 1.0 that a system typically then maps to a class prediction.

Not every model that outputs numerical predictions is a regression model. In some cases, a numeric prediction is really just a classification model that happens to have numeric class names. For example, a model that predicts a numeric postal code is a classification model, not a regression model.

নিয়মিতকরণ

#fundamentals

Any mechanism that reduces overfitting . Popular types of regularization include:

Regularization can also be defined as the penalty on a model's complexity.

regularization rate

#fundamentals

A number that specifies the relative importance of regularization during training. Raising the regularization rate reduces overfitting but may reduce the model's predictive power. Conversely, reducing or omitting the regularization rate increases overfitting.

ReLU

#fundamentals

Abbreviation for Rectified Linear Unit .

retrieval-augmented generation

#fundamentals

A software architecture commonly used in large language model (LLM) applications. Common motivations to use retrieval-augmented generation include:

  • Increasing the factual accuracy of the model's generated responses
  • Giving the model access to knowledge it was not trained on
  • Changing what knowledge the model uses
  • Enabling the model to cite sources

For example, suppose that a chemistry app uses the PaLM API to generate summaries related to user queries. When the app's backend receives a query, the backend first searches for ("retrieves") data that's relevant to the user's query, appends ("augments") the relevant chemistry data to the user's query, and instructs the LLM to create a summary based on the appended data.

ROC (receiver operating characteristic) Curve

#fundamentals

A graph of true positive rate vs. false positive rate for different classification thresholds in binary classification.

The shape of an ROC curve suggests a binary classification model's ability to separate positive classes from negative classes. Suppose, for example, that a binary classification model perfectly separates all the negative classes from all the positive classes:

A number line with 8 positive examples on the right side and
          7 negative examples on the left.

The ROC curve for the preceding model looks as follows:

An ROC curve. The x-axis is False Positive Rate and the y-axis
          is True Positive Rate. The curve has an inverted L shape. The curve
          starts at (0.0,0.0) and goes straight up to (0.0,1.0). Then the curve
          goes from (0.0,1.0) to (1.0,1.0).

In contrast, the following illustration graphs the raw logistic regression values for a terrible model that can't separate negative classes from positive classes at all:

A number line with positive examples and negative classes
          completely intermixed.

The ROC curve for this model looks as follows:

An ROC curve, which is actually a straight line from (0.0,0.0)
          to (1.0,1.0).

Meanwhile, back in the real world, most binary classification models separate positive and negative classes to some degree, but usually not perfectly. So, a typical ROC curve falls somewhere between the two extremes:

An ROC curve. The x-axis is False Positive Rate and the y-axis
          is True Positive Rate. The ROC curve approximates a shaky arc
          traversing the compass points from West to North.

The point on an ROC curve closest to (0.0,1.0) theoretically identifies the ideal classification threshold. However, several other real-world issues influence the selection of the ideal classification threshold. For example, perhaps false negatives cause far more pain than false positives.

A numerical metric called AUC summarizes the ROC curve into a single floating-point value.

রুট গড় বর্গাকার ত্রুটি (RMSE)

#fundamentals

The square root of the Mean Squared Error .

এস

sigmoid function

#fundamentals

A mathematical function that "squishes" an input value into a constrained range, typically 0 to 1 or -1 to +1. That is, you can pass any number (two, a million, negative billion, whatever) to a sigmoid and the output will still be in the constrained range. A plot of the sigmoid activation function looks as follows:

A two-dimensional curved plot with x values spanning the domain
          -infinity to +positive, while y values span the range almost 0 to
          almost 1. When x is 0, y is 0.5. The slope of the curve is always
          positive, with the highest slope at 0,0.5 and gradually decreasing
          slopes as the absolute value of x increases.

The sigmoid function has several uses in machine learning, including:

softmax

#fundamentals

A function that determines probabilities for each possible class in a multi-class classification model . The probabilities add up to exactly 1.0. For example, the following table shows how softmax distributes various probabilities:

Image is a... সম্ভাবনা
কুকুর .85
বিড়াল .13
ঘোড়া .02

Softmax is also called full softmax .

Contrast with candidate sampling .

sparse feature

#ভাষা
#fundamentals

A feature whose values are predominately zero or empty. For example, a feature containing a single 1 value and a million 0 values is sparse. In contrast, a dense feature has values that are predominantly not zero or empty.

In machine learning, a surprising number of features are sparse features. Categorical features are usually sparse features. For example, of the 300 possible tree species in a forest, a single example might identify just a maple tree . Or, of the millions of possible videos in a video library, a single example might identify just "Casablanca."

In a model, you typically represent sparse features with one-hot encoding . If the one-hot encoding is big, you might put an embedding layer on top of the one-hot encoding for greater efficiency.

sparse representation

#ভাষা
#fundamentals

Storing only the position(s) of nonzero elements in a sparse feature.

For example, suppose a categorical feature named species identifies the 36 tree species in a particular forest. Further assume that each example identifies only a single species.

You could use a one-hot vector to represent the tree species in each example. A one-hot vector would contain a single 1 (to represent the particular tree species in that example) and 35 0 s (to represent the 35 tree species not in that example). So, the one-hot representation of maple might look something like the following:

A vector in which positions 0 through 23 hold the value 0, position
          24 holds the value 1, and positions 25 through 35 hold the value 0.

Alternatively, sparse representation would simply identify the position of the particular species. If maple is at position 24, then the sparse representation of maple would simply be:

24

Notice that the sparse representation is much more compact than the one-hot representation.

sparse vector

#fundamentals

A vector whose values are mostly zeroes. See also sparse feature and sparsity .

squared loss

#fundamentals

Synonym for L 2 loss .

স্থির

#fundamentals

Something done once rather than continuously. The terms static and offline are synonyms. The following are common uses of static and offline in machine learning:

  • static model (or offline model ) is a model trained once and then used for a while.
  • static training (or offline training ) is the process of training a static model.
  • static inference (or offline inference ) is a process in which a model generates a batch of predictions at a time.

Contrast with dynamic .

static inference

#fundamentals

Synonym for offline inference .

স্থিরতা

#fundamentals

A feature whose values don't change across one or more dimensions, usually time. For example, a feature whose values look about the same in 2021 and 2023 exhibits stationarity.

In the real world, very few features exhibit stationarity. Even features synonymous with stability (like sea level) change over time.

Contrast with nonstationarity .

stochastic gradient descent (SGD)

#fundamentals

A gradient descent algorithm in which the batch size is one. In other words, SGD trains on a single example chosen uniformly at random from a training set .

supervised machine learning

#fundamentals

Training a model from features and their corresponding labels . Supervised machine learning is analogous to learning a subject by studying a set of questions and their corresponding answers. After mastering the mapping between questions and answers, a student can then provide answers to new (never-before-seen) questions on the same topic.

Compare with unsupervised machine learning .

synthetic feature

#fundamentals

A feature not present among the input features, but assembled from one or more of them. Methods for creating synthetic features include the following:

  • Bucketing a continuous feature into range bins.
  • Creating a feature cross .
  • Multiplying (or dividing) one feature value by other feature value(s) or by itself. For example, if a and b are input features, then the following are examples of synthetic features:
    • ab
    • একটি 2
  • Applying a transcendental function to a feature value. For example, if c is an input feature, then the following are examples of synthetic features:
    • sin(c)
    • ln(c)

Features created by normalizing or scaling alone are not considered synthetic features.

টি

test loss

#fundamentals

A metric representing a model's loss against the test set . When building a model , you typically try to minimize test loss. That's because a low test loss is a stronger quality signal than a low training loss or low validation loss .

A large gap between test loss and training loss or validation loss sometimes suggests that you need to increase the regularization rate .

প্রশিক্ষণ

#fundamentals

The process of determining the ideal parameters (weights and biases) comprising a model . During training, a system reads in examples and gradually adjusts parameters. Training uses each example anywhere from a few times to billions of times.

training loss

#fundamentals

A metric representing a model's loss during a particular training iteration. For example, suppose the loss function is Mean Squared Error . Perhaps the training loss (the Mean Squared Error) for the 10th iteration is 2.2, and the training loss for the 100th iteration is 1.9.

A loss curve plots training loss vs. the number of iterations. A loss curve provides the following hints about training:

  • A downward slope implies that the model is improving.
  • An upward slope implies that the model is getting worse.
  • A flat slope implies that the model has reached convergence .

For example, the following somewhat idealized loss curve shows:

  • A steep downward slope during the initial iterations, which implies rapid model improvement.
  • A gradually flattening (but still downward) slope until close to the end of training, which implies continued model improvement at a somewhat slower pace then during the initial iterations.
  • A flat slope towards the end of training, which suggests convergence.

The plot of training loss vs. iterations. This loss curve starts
     with a steep downward slope. The slope gradually flattens until the
     slope becomes zero.

Although training loss is important, see also generalization .

training-serving skew

#fundamentals

The difference between a model's performance during training and that same model's performance during serving .

প্রশিক্ষণ সেট

#fundamentals

The subset of the dataset used to train a model .

Traditionally, examples in the dataset are divided into the following three distinct subsets:

Ideally, each example in the dataset should belong to only one of the preceding subsets. For example, a single example should not belong to both the training set and the validation set.

সত্য নেতিবাচক (TN)

#fundamentals

An example in which the model correctly predicts the negative class . For example, the model infers that a particular email message is not spam , and that email message really is not spam .

সত্য ইতিবাচক (TP)

#fundamentals

An example in which the model correctly predicts the positive class . For example, the model infers that a particular email message is spam, and that email message really is spam.

true positive rate (TPR)

#fundamentals

Synonym for recall . এটাই:

$$\text{true positive rate} = \frac{\text{true positives}} {\text{true positives} + \text{false negatives}}$$

True positive rate is the y-axis in an ROC curve .

underfitting

#fundamentals

Producing a model with poor predictive ability because the model hasn't fully captured the complexity of the training data. Many problems can cause underfitting, including:

unlabeled example

#fundamentals

An example that contains features but no label . For example, the following table shows three unlabeled examples from a house valuation model, each with three features but no house value:

বেডরুমের সংখ্যা বাথরুমের সংখ্যা House age
3 2 15
2 1 72
4 2 34

In supervised machine learning , models train on labeled examples and make predictions on unlabeled examples .

In semi-supervised and unsupervised learning, unlabeled examples are used during training.

Contrast unlabeled example with labeled example .

unsupervised machine learning

#clustering
#fundamentals

Training a model to find patterns in a dataset, typically an unlabeled dataset.

The most common use of unsupervised machine learning is to cluster data into groups of similar examples. For example, an unsupervised machine learning algorithm can cluster songs based on various properties of the music. The resulting clusters can become an input to other machine learning algorithms (for example, to a music recommendation service). Clustering can help when useful labels are scarce or absent. For example, in domains such as anti-abuse and fraud, clusters can help humans better understand the data.

Contrast with supervised machine learning .

ভি

বৈধতা

#fundamentals

The initial evaluation of a model's quality. Validation checks the quality of a model's predictions against the validation set .

Because the validation set differs from the training set , validation helps guard against overfitting .

You might think of evaluating the model against the validation set as the first round of testing and evaluating the model against the test set as the second round of testing.

validation loss

#fundamentals

A metric representing a model's loss on the validation set during a particular iteration of training.

See also generalization curve .

validation set

#fundamentals

The subset of the dataset that performs initial evaluation against a trained model . Typically, you evaluate the trained model against the validation set several times before evaluating the model against the test set .

Traditionally, you divide the examples in the dataset into the following three distinct subsets:

Ideally, each example in the dataset should belong to only one of the preceding subsets. For example, a single example should not belong to both the training set and the validation set.

ডব্লিউ

ওজন

#fundamentals

A value that a model multiplies by another value. Training is the process of determining a model's ideal weights; inference is the process of using those learned weights to make predictions.

weighted sum

#fundamentals

The sum of all the relevant input values multiplied by their corresponding weights. For example, suppose the relevant inputs consist of the following:

ইনপুট মান input weight
2 -1.3
-1 0.6
3 0.4

The weighted sum is therefore:

weighted sum = (2)(-1.3) + (-1)(0.6) + (3)(0.4) = -2.0

A weighted sum is the input argument to an activation function .

জেড

Z-score normalization

#fundamentals

A scaling technique that replaces a raw feature value with a floating-point value representing the number of standard deviations from that feature's mean. For example, consider a feature whose mean is 800 and whose standard deviation is 100. The following table shows how Z-score normalization would map the raw value to its Z-score:

কাঁচা মান Z-score
800 0
950 +1.5
575 -2.25

The machine learning model then trains on the Z-scores for that feature instead of on the raw values.