{"id":195,"date":"2024-03-26T07:49:10","date_gmt":"2024-03-26T07:49:10","guid":{"rendered":"https:\/\/shatranj.ai\/?page_id=195"},"modified":"2026-02-27T03:28:27","modified_gmt":"2026-02-27T03:28:27","slug":"temat-e-kurrikules","status":"publish","type":"page","link":"https:\/\/shatranj.ai\/sq\/temat-e-kurrikules\/","title":{"rendered":"Temat e kurrikul\u00ebs"},"content":{"rendered":"
\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Temat e kurrikul\u00ebs<\/h1>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Kurrikula e projektit Shatranj.ai mund t\u00eb aksesohet p\u00ebrmes lms.shatranj.ai<\/a><\/strong><\/p>

M\u00eb posht\u00eb jan\u00eb p\u00ebrmbledhjet e shkurtra t\u00eb temave brenda kurrikul\u00ebs.<\/p>

M\u00ebsimi 1 \u2013 Fush\u00ebveprimi i kursit dhe prioritetet<\/h1>
  • Prezanton projektin Shatranj.AI, themelet e tij Erasmus+, organizatat partnere dhe platformat digjitale.<\/li>
  • Vizioni i projektit, konteksti i Erasmus KA2<\/li>
  • Institucionet partnere dhe fokusi n\u00eb trash\u00ebgimin\u00eb kulturore<\/li>
  • P\u00ebrmbledhje e platformave (redaktori, LMS, mjetet e kodimit)<\/li>
  • Rol\u00ebt e m\u00ebsuesve dhe rezultatet e nx\u00ebn\u00ebsve<\/li>
  • P\u00ebrmbledhje e struktur\u00ebs s\u00eb kurrikul\u00ebs<\/li>
  • Hyrje n\u00eb Python\/Jupyter<\/li><\/ul>

    M\u00ebsimi 2 \u2013 Hyrje n\u00eb Informatik\u00eb dhe Konfigurimi i Python-it<\/h1>
    • Student\u00ebt m\u00ebsojn\u00eb konceptet themelore t\u00eb p\u00ebrpunimit t\u00eb t\u00eb dh\u00ebnave dhe instalojn\u00eb Python\/Jupyter.<\/li>
    • CPU, RAM, I\/O baz\u00eb<\/li>
    • Bit\u00eb, bajtet, p\u00ebrfaq\u00ebsimi binar<\/li>
    • Instalimi i JupyterLab<\/li>
    • Ekzekutimi i par\u00eb i bllokut Python<\/li>
    • Variablat, shprehjet e thjeshta<\/li>
    • Qasja n\u00eb dosjet Drive<\/li><\/ul>

      M\u00ebsimi 3 \u2013 Llojet e t\u00eb dh\u00ebnave n\u00eb Python<\/h1>
      • P\u00ebrfshin llojet e t\u00eb dh\u00ebnave t\u00eb integruara t\u00eb Python-it dhe operacionet baz\u00eb.<\/li>
      • Integer\u00eb, numra me presje, vargje, boolean\u00eb<\/li>
      • Konvertimi i tipit<\/li>
      • Lista dhe indeksimi<\/li>
      • Konceptet e ndryshueshm\u00ebris\u00eb<\/li>
      • Ushtrime me figurat e shahut si fije<\/li><\/ul>

        M\u00ebsimi 4 \u2013 Kushtet, ciklet, rrjedha e kontrollit<\/h1>
        • Prezanton logjik\u00ebn, ciklet dhe programet interaktive.<\/li>
        • Logjika N\u00ebse\/elif\/tjet\u00ebr<\/li>
        • Operacionet Booleane<\/li>
        • Ciklet For\/While<\/li>
        • Ndalo\/vazhdo<\/li>
        • Programet e thjeshta t\u00eb hyrjes<\/li><\/ul>

          M\u00ebsimi 5 \u2013 Funksionet, Fush\u00ebveprimi, Parametrat<\/h1>
          • M\u00ebson kod modular me funksione.<\/li>
          • P\u00ebrcaktimi i funksioneve<\/li>
          • Parametrat dhe kthimet<\/li>
          • Fusha lokale\/globale<\/li>
          • Lambda-t\u00eb<\/li>
          • Projekt funksional i vog\u00ebl (llogarit\u00ebs i vler\u00ebs s\u00eb pjes\u00ebs)<\/li><\/ul>

            M\u00ebsimi 6 \u2013 Skedar\u00eb, p\u00ebrjashtime, biblioteka, testimi<\/h1>
            • Puna me skedar\u00eb dhe kod t\u00eb q\u00ebndruesh\u00ebm.<\/li>
            • Skedar lexo\/shkruaj<\/li>
            • P\u00ebrpiqu\/p\u00ebrjashto<\/li>
            • Importimi i bibliotekave<\/li>
            • Testim i thjesht\u00eb<\/li>
            • Menaxhimi i hyrjeve t\u00eb pavlefshme<\/li><\/ul>

              M\u00ebsimi 7 \u2013 OOP, Klasat, TicTacToe<\/h1>
              • Ekspozimi i par\u00eb me OOP.<\/li>
              • Klasat dhe objektet<\/li>
              • Atributet dhe metodat<\/li>
              • Modelimin e loj\u00ebrave<\/li>
              • Zbatimi i TicTacToe<\/li>
              • Debugimi i kodit OOP<\/li><\/ul>

                M\u00ebsimet 8 \u2013 P\u00ebrfaq\u00ebsimi i tabel\u00ebs s\u00eb shahut dhe shatranjit<\/h1>
                • P\u00ebrfaq\u00ebsimet e tabel\u00ebs p\u00ebr shah dhe shatranxh<\/li>
                • Sistemet e koordinatave dhe strategjit\u00eb e indeksimit<\/li>
                • Strukturat e t\u00eb dh\u00ebnave t\u00eb brendshme p\u00ebr gjendjen e bordit<\/li>
                • UTF-8 dhe paraqitja simbolike e pjes\u00ebve<\/li>
                • Integrimi me redaktor\u00ebt e bordit dhe mjetet e vizualizimit<\/li><\/ul>

                  M\u00ebsimet 9 \u2013 L\u00ebvizja e copave, P\u00ebrdit\u00ebsimet e gjendjes s\u00eb loj\u00ebs dhe Kushtet terminale<\/h1>
                  • Rregullat e l\u00ebvizjes s\u00eb figurave n\u00eb shah dhe shatranxh<\/li>
                  • Gjenerimi i l\u00ebvizjeve ligjore kundrejt atyre pseudo-ligjore<\/li>
                  • P\u00ebrdit\u00ebsimi i gjendjes s\u00eb loj\u00ebs pas nj\u00eb l\u00ebvizjeje<\/li>
                  • Zbulimi i kontrollit dhe vet\u00ebkontrollit t\u00eb paligjsh\u00ebm<\/li>
                  • Detektimi i kushteve terminale: mat dhe pat<\/li><\/ul>

                    M\u00ebsimet 10 \u2013 Problemet e k\u00ebrkimit dhe kalimi n\u00eb graf<\/h1>
                    • Formulimi i problemit t\u00eb k\u00ebrkimit: gjendjet, veprimet, tranzicionet dhe q\u00ebllimet<\/li>
                    • Graf\u00ebt dhe pem\u00ebt e hap\u00ebsir\u00ebs s\u00eb gjendjes<\/li>
                    • K\u00ebrkimi n\u00eb thell\u00ebsi t\u00eb par\u00eb (DFS)<\/li>
                    • K\u00ebrkimi me gjer\u00ebsi t\u00eb par\u00eb (BFS)<\/li>
                    • K\u00ebrkimi me kosto uniforme (UCS)<\/li>
                    • Udh\u00ebzime p\u00ebr gjurmimin e graf\u00ebve dhe vizualizim<\/li>
                    • Shembuj t\u00eb thjesht\u00eb me shah dhe t\u00eb bazuar n\u00eb rrjet\u00eb<\/li><\/ul>

                      Ligj\u00ebratat 11 \u2013 K\u00ebrkimi heuristik dhe pem\u00ebt e loj\u00ebs kund\u00ebrshtare<\/h1>
                      • Funksionet heuristike dhe k\u00ebrkimi i informuar<\/li>
                      • Pranueshm\u00ebria dhe q\u00ebndrueshm\u00ebria<\/li>
                      • K\u00ebrkim A*<\/li>
                      • Pema e k\u00ebrkimit dhe e loj\u00ebs kund\u00ebrshtare<\/li>
                      • Funksionet e vler\u00ebsimit p\u00ebr gjendjet e loj\u00ebs<\/li>
                      • K\u00ebrkimi minimax<\/li>
                      • Expectiminimax p\u00ebr mjedise stocastike dhe t\u00eb pasigurta<\/li>
                      • Prunimi Alpha\u2013beta dhe p\u00ebrmir\u00ebsimet e performanc\u00ebs<\/li>
                      • Funksionet e vler\u00ebsimit p\u00ebr gjendjet e loj\u00ebs<\/li>
                      • Shembuj armiq\u00ebsor\u00eb t\u00eb bazuar n\u00eb shah<\/li><\/ul>

                        M\u00ebsimi 12 \u2013 Tura me kal\u00eb (Tura e kalor\u00ebsit)<\/h1>
                        • Eksploron turneun e kalor\u00ebsit me rikursion dhe heuristika.<\/li>
                        • L\u00ebvizja e grafit kalor\u00ebs<\/li>
                        • Turne t\u00eb hapura\/t\u00eb mbyllura<\/li>
                        • Kthim mbrapa me DFS<\/li>
                        • Euristika Warnsdorff<\/li>
                        • Lidhje me TSP<\/li><\/ul>

                          M\u00ebsimi 13 \u2013 Enigma e tet\u00eb mbret\u00ebreshave<\/h1>
                          • K\u00ebnaqja e kufizimeve me rikthim prapa.<\/li>
                          • Logjika e sulmit t\u00eb Mbret\u00ebresh\u00ebs<\/li>
                          • K\u00ebrkim rekursiv<\/li>
                          • Teknikat e optimizimit<\/li>
                          • Referenca historike p\u00ebr mbret\u00ebresh\u00ebn<\/li>
                          • Zbatimet e bllokut t\u00eb sh\u00ebnimeve<\/li><\/ul>

                            M\u00ebsimi 14 \u2013 Gruri dhe problemi i tabel\u00ebs s\u00eb shahut<\/h1>
                            • Hangavere matematikore dhe rritje eksponenciale.<\/li>
                            • Dyfishimi n\u00eb tabel\u00ebn e shahut<\/li>
                            • Fuqit\u00eb e 2-s\u00eb<\/li>
                            • Enigma lloj-Mount Fuji t\u00eb p\u00ebrdorura n\u00eb intervista, enigma t\u00eb matematik\u00ebs s\u00eb aplikuar<\/li>
                            • Katrore magjike<\/li>
                            • Hangav\u00ebrtit\u00eb logjike t\u00eb Smullyanit<\/li>
                            • Problemet me pllakat Knight dhe problemet e tjera me pllakat n\u00eb rrjet\u00eb<\/li><\/ul>

                              M\u00ebsimi 15 \u2013 Minimax, Alpha-Beta, Logika e Matjes s\u00eb Mbretit<\/h1>
                              • K\u00ebrkim i thell\u00eb kund\u00ebrshtar dhe fundloj\u00ebra shahu.<\/li>
                              • Llogaritje minimax<\/li>
                              • Prerja alfa-beta<\/li>
                              • Opozit\u00eb, triangulim<\/li>
                              • Burime historike (Al-Adli, Reti)<\/li><\/ul>

                                M\u00ebsimi 16 \u2013 Diamanti i Sulit (Studim historik i fundloj\u00ebs)<\/h1>
                                • Analiza historike e fundloj\u00ebs s\u00eb shahut, bazat e t\u00eb dh\u00ebnave t\u00eb fundloj\u00ebs, programimi dinamik, hashing<\/li>
                                • Biografia e Al-Suli<\/li>
                                • Rikonstruimi i faz\u00ebs p\u00ebrfundimtare<\/li>
                                • Opozita dhe triangulimi<\/li>
                                • Teoria e katror\u00ebve p\u00ebrkat\u00ebs<\/li>
                                • Zgjidhja me kod e programimit dinamik p\u00ebr dilem\u00ebn 1000-vje\u00e7are n\u00eb C\/C++<\/li>
                                • Inspektimi i zgjidhjes n\u00eb play.shatranj.ai dhe gjithashtu p\u00ebrmes blloqeve t\u00eb sh\u00ebnimeve dhe tabelave ASCII<\/li><\/ul>

                                  M\u00ebsimi 17 \u2013 P\u00ebrshtatja e Stockfish p\u00ebr t\u00eb luajtur Shatranj, Rybka\u2013Deep Blue\u2013Stockfish Historia<\/h1>

                                  Eksploron se si motor\u00ebt modern\u00eb t\u00eb shahut kan\u00eb evoluar dhe si motor\u00ebt me kod t\u00eb hapur mund t\u00eb p\u00ebrshtaten me variantet historike.<\/p>

                                  • K\u00ebrkimi me forc\u00eb bruto i Deep Blue<\/li>
                                  • Rybka dhe ngritja e motor\u00ebve t\u00eb p\u00ebrqendruar n\u00eb vler\u00ebsim<\/li>
                                  • Stockfish si nj\u00eb motor i hapur, i drejtuar nga komuniteti<\/li>
                                  • Si Stockfish p\u00ebrfaq\u00ebson figurat, l\u00ebvizjet dhe rregullat<\/li>
                                  • Modifikimi i l\u00ebvizjes s\u00eb figurave (ferz, vezir), vler\u00ebsimi, rregullat e ligjshm\u00ebris\u00eb<\/li>
                                  • Nd\u00ebrtimi i k\u00ebrkimit dhe vler\u00ebsimit t\u00eb p\u00ebrputhsh\u00ebm me Shatranjin<\/li><\/ul>

                                    M\u00ebsimi 18 \u2013 Bazat e M\u00ebsimit t\u00eb P\u00ebrforcimit: Bot\u00eb rrjetore, Programim dinamik dhe Kompleksitet<\/h1>

                                    Prezanton m\u00ebsimin e p\u00ebrforcimit (RL) duke zgjidhur nj\u00eb bot\u00eb rrjet\u00eb t\u00eb vog\u00ebl sakt\u00ebsisht kur rregullat jan\u00eb t\u00eb njohura, pastaj tregon pse ky qasje \u201ci gjithdijsh\u00ebm\u201d d\u00ebshton n\u00eb loj\u00ebra t\u00eb m\u00ebdha si shahu.<\/p>

                                    • Cikli agjent\u2013mjedis; gjendjet, veprimet, shp\u00ebrblimet, episodet; faktori i zbritjes \u03b3.<\/li>
                                    • Vler\u00ebsimi i politik\u00ebs (\u201crobot i rr\u00ebshqitsh\u00ebm\u201d) dhe iteracioni i vler\u00ebs (\u201cgjuetar i thesarit\u201d) duke p\u00ebrdorur rezervimet e Bellmanit.<\/li>
                                    • P\u00ebrhapja e vler\u00ebs vizuale dhe nxjerrja e nj\u00eb politike optimale nga funksioni i vler\u00ebs.<\/li>
                                    • Mallkimi i dimensionalitetit: kompleksiteti i hap\u00ebsir\u00ebs s\u00eb gjendjes kundrejt atij t\u00eb pem\u00ebs s\u00eb loj\u00ebs; motivimi i numrit t\u00eb Shannon-it.<\/li>
                                    • Loj\u00ebrat historike \u201cgjigande\u201d (p.sh., shahu i Tamerlanit, Go) si kontekst p\u00ebrse \u00ebsht\u00eb e nevojshme t\u00eb m\u00ebsohet.<\/li><\/ul>

                                      M\u00ebsimi 19 \u2013 Mbret\u00ebresha e Ngrir\u00eb: M\u00ebsimi Q n\u00eb tabel\u00eb n\u00eb FrozenLake<\/h1>

                                      L\u00ebvizjet nga planifikimi drejt m\u00ebsimit: agjenti fillon pa hart\u00eb dhe m\u00ebson nj\u00eb politik\u00eb p\u00ebrmes prov\u00ebs dhe gabimit duke p\u00ebrdorur Q-learning tabelar.<\/p>

                                      • Formuloni FrozenLake\/Frozen Rook si nj\u00eb MDP: S, A, R, P, gjendjet terminale, \u03b3.<\/li>
                                      • Rregulli i p\u00ebrdit\u00ebsimit t\u00eb Q-learning dhe eksplorimi \u03b5-greedy (orari eksplorim\u2192shfryt\u00ebzim).<\/li>
                                      • Trajno nj\u00eb agjent n\u00eb Gymnasium FrozenLake; krahaso tranzicionet deterministike me ato rr\u00ebshqit\u00ebse.<\/li>
                                      • Inspektoni at\u00eb q\u00eb \u00ebsht\u00eb m\u00ebsuar p\u00ebrmes hartave t\u00eb nxeht\u00ebsis\u00eb n\u00eb Q-tabela \/ shigjetave t\u00eb politik\u00ebs; rregulloni \u03b1, \u03b3, \u03b5 dhe numrin e episodeve.<\/li>
                                      • M\u00ebsimet e shkall\u00ebzimit: shp\u00ebrblime t\u00eb rralla, njohje e vonuar dhe pse hartat m\u00eb t\u00eb m\u00ebdha jan\u00eb m\u00eb t\u00eb v\u00ebshtira.<\/li><\/ul>

                                        M\u00ebsimi 20 \u2013 Mat me dy turre kund\u00ebr mbretit t\u00eb vet\u00ebm: M\u00ebsimi me diferenc\u00eb kohore n\u00eb praktik\u00eb<\/h1>

                                        Aplikon Q-learning n\u00eb nj\u00eb fund loje t\u00eb vog\u00ebl shahu dhe e b\u00ebn baz\u00ebn e kodit RL \u201ct\u00eb v\u00ebrtet\u00eb\u201d duke ndar\u00eb bllokun e sh\u00ebnimeve t\u00eb eksperimentit nga modulet e t\u00eb m\u00ebsuarit dhe trajnimit.<\/p>

                                        • M\u00ebsimi me ndryshim kohor (TD): identifikoni gabimin TD brenda p\u00ebrdit\u00ebsimit t\u00eb Q-learning; pse p\u00ebrdit\u00ebsimet TD gjat\u00eb loj\u00ebs.<\/li>
                                        • Pse m\u00ebsimi Monte Carlo \u00ebsht\u00eb shum\u00eb i ngadalt\u00eb p\u00ebr loj\u00ebra t\u00eb ngjashme me shahun me shp\u00ebrblim t\u00eb vonuar.<\/li>
                                        • Staku i inxhinieris\u00eb: rl.py (Q-memory + p\u00ebrdit\u00ebsimi TD), trainer.py (cikli i episodit, orari i eksplorimit), notebook si laborator.<\/li>
                                        • Kodoni pozicionet e shahut si gjendje t\u00eb lexueshme nga makina (FEN) dhe trajnoni nj\u00eb agjent tabelar n\u00eb nj\u00eb hap\u00ebsir\u00eb gjendjesh t\u00eb kufizuara t\u00eb finales s\u00eb loj\u00ebs\/enigm\u00ebs.<\/li>
                                        • Kufizimet: pse metodat tabulare d\u00ebshtojn\u00eb p\u00ebr shahun e plot\u00eb (mallkimi i dimensionalitetit) dhe nevoja p\u00ebr afrimimin e funksioneve.<\/li><\/ul>

                                          M\u00ebsimi 21 \u2013 Rrjetet e thella Q: nga tabelat Q te rrjetet nervore<\/h1>

                                          Prezanton afrimin e funksionit p\u00ebr RL duke z\u00ebvend\u00ebsuar tabel\u00ebn Q me nj\u00eb rrjet\u00eb nervore (DQN) dhe duke e aplikuar at\u00eb n\u00eb disa loj\u00ebra tavoline t\u00eb vogla.<\/p>

                                          • Pse tabelat Q nuk shkall\u00ebzohen: ka shum\u00eb gjendje; p\u00ebrgjith\u00ebsimi k\u00ebrkon nj\u00eb model q\u00eb mund t\u00eb \u201cparashikoj\u00eb\u201d vlera p\u00ebr pozicione t\u00eb panjohura.<\/li>
                                          • Cikli i trajnimit t\u00eb Deep Q-Network (DQN): buferi i riprodhimit, rrjeti target, p\u00ebrdit\u00ebsimet me mini-batch, zbehja \u03b5.<\/li>
                                          • Zbato dhe eksperimento me DQN n\u00eb loj\u00ebra si Connect-4 (4Connect), Fox & Hounds dhe Othello\/Reversi.<\/li>
                                          • Diagnostika: kurbat e t\u00eb m\u00ebsuarit, \u00e7\u00ebshtjet e stabilitetit (mbitestimimi, divergjenca) dhe masat praktike p\u00ebr zbutje.<\/li>
                                          • Krahaso qasjet: DQN kundrejt vler\u00ebsimit n\u00eb stil NNUE dhe vler\u00ebsimit t\u00eb hartuar me dor\u00eb (HCE) p\u00ebr t\u00eb diskutuar kompromiset e arkitektur\u00ebs.<\/li><\/ul>

                                            M\u00ebsimi 22 \u2013 Monte Carlo Rollouts dhe MCTS n\u00eb Qirkat<\/h1>

                                            Nd\u00ebrton nj\u00eb mjedis t\u00eb plot\u00eb Qirkat dhe m\u00eb pas p\u00ebrparon nga implementime t\u00eb rast\u00ebsishme drejt nj\u00eb k\u00ebrkimi t\u00eb plot\u00eb t\u00eb pem\u00ebs Monte Carlo (MCTS) me selektim UCT.<\/p>

                                            • Zbato korniz\u00ebn e rregullave Qirkat (tabela 5\u00d75, C3 bosh) dhe rregull\u00ebn e kapjes maksimale q\u00eb detyron sekuenca kapjesh.<\/li>
                                            • Gjenerimi i l\u00ebvizjes q\u00eb num\u00ebron linjat e kapjes, imponon kapjen e detyrueshme dhe filtron kapjet me gjat\u00ebsi maksimale.<\/li>
                                            • Vijat baz\u00eb t\u00eb Monte Carlos: shp\u00ebrndarje t\u00eb rast\u00ebsishme dhe vler\u00ebsim i shesht\u00eb i l\u00ebvizjeve Monte Carlo para shtimit t\u00eb ri-p\u00ebrdorimit t\u00eb pem\u00ebs.<\/li>
                                            • Pipeline-i MCTS: seleksionim, zgjerim, zbatim\/vler\u00ebsim, p\u00ebrhapje e prapme; zgjedhja p\u00ebrfundimtare e l\u00ebvizjes me UCT\/num\u00ebr vizitash.<\/li>
                                            • Regjistrime t\u00eb riprodhueshme t\u00eb loj\u00ebs dhe mjete auditimi p\u00ebr riprodhim hap pas hapi dhe debugim.<\/li><\/ul>

                                              M\u00ebsimi 23 \u2013 AlphaZero n\u00eb Othello\/Reversi<\/h1>

                                              P\u00ebrmir\u00ebson MCTS-n\u00eb n\u00eb nj\u00eb k\u00ebrkim n\u00eb stilin AlphaZero duke shtuar nj\u00eb rrjet nervor q\u00eb furnizon nj\u00eb politik\u00eb paraprake dhe nj\u00eb vler\u00ebsim t\u00eb vler\u00ebs, pastaj st\u00ebrvitet p\u00ebrmes vet\u00eb-loj\u00ebs.<\/p>

                                              • Urdh\u00ebroni intuit\u00ebn me nj\u00eb demonstrim t\u00eb vog\u00ebl \u2018Connect2\u2019 AlphaZero, pastaj transferoni idet\u00eb n\u00eb Othello.<\/li>
                                              • Z\u00ebvend\u00ebsoni UCT me PUCT: kombinoni statistikat e vizitave me nj\u00eb politik\u00eb t\u00eb m\u00ebsuar paraprakisht p\u00ebr t\u00eb udh\u00ebhequr eksplorimin.<\/li>
                                              • Kokat e rrjetit nervor: politika (probabilitetet e l\u00ebvizjeve) dhe vlera (vler\u00ebsimi i pozicionit) p\u00ebrdoren n\u00eb vend t\u00eb simulimeve t\u00eb rast\u00ebsishme.<\/li>
                                              • Cikli AlphaZero: vet\u00eb-loj\u00eb \u2192 objektivat e trajnimit (\u03c0, z) \u2192 p\u00ebrdit\u00ebsim i rrjetit \u2192 p\u00ebrs\u00ebrit; vler\u00ebsim p\u00ebrmes ndeshjeve\/log-eve t\u00eb turneut.<\/li>
                                              • Kodimi i l\u00ebvizjeve i nd\u00ebrgjegjsh\u00ebm ndaj rrug\u00ebs p\u00ebr sekuenca kapjeje me gjat\u00ebsi t\u00eb ndryshueshme, n\u00eb m\u00ebnyr\u00eb q\u00eb rrug\u00ebt e ndryshme t\u00eb kapjes t\u00eb mbeten t\u00eb dallueshme.<\/li><\/ul>

                                                M\u00ebsimi 24 \u2013 AlphaZero n\u00eb Qirkat: PUCT, rrjetat e politik\u00ebs\/vler\u00ebs dhe vet\u00eb-loja<\/h1>

                                                P\u00ebrmir\u00ebson MCTS-n\u00eb n\u00eb nj\u00eb k\u00ebrkim n\u00eb stilin AlphaZero duke shtuar nj\u00eb rrjet nervor q\u00eb furnizon nj\u00eb politik\u00eb paraprake dhe nj\u00eb vler\u00ebsim t\u00eb vler\u00ebs, pastaj st\u00ebrvitet p\u00ebrmes vet\u00eb-loj\u00ebs.<\/p>

                                                • Urdh\u00ebro intuit\u00ebn me nj\u00eb demonstrim t\u00eb vog\u00ebl \u2018Connect2\u2019 AlphaZero, pastaj transfero idet\u00eb te Qirkat.<\/li>
                                                • Z\u00ebvend\u00ebsoni UCT me PUCT: kombinoni statistikat e vizitave me nj\u00eb politik\u00eb t\u00eb m\u00ebsuar paraprakisht p\u00ebr t\u00eb udh\u00ebhequr eksplorimin.<\/li>
                                                • Kokat e rrjetit nervor: politika (probabilitetet e l\u00ebvizjeve) dhe vlera (vler\u00ebsimi i pozicionit) p\u00ebrdoren n\u00eb vend t\u00eb simulimeve t\u00eb rast\u00ebsishme.<\/li>
                                                • Cikli AlphaZero: vet\u00eb-loj\u00eb \u2192 objektivat e trajnimit (\u03c0, z) \u2192 p\u00ebrdit\u00ebsim i rrjetit \u2192 p\u00ebrs\u00ebrit; vler\u00ebsim p\u00ebrmes ndeshjeve\/log-eve t\u00eb turneut.<\/li>
                                                • Kodimi i l\u00ebvizjeve i nd\u00ebrgjegjsh\u00ebm ndaj rrug\u00ebs p\u00ebr sekuenca kapjeje me gjat\u00ebsi t\u00eb ndryshueshme, n\u00eb m\u00ebnyr\u00eb q\u00eb rrug\u00ebt e ndryshme t\u00eb kapjes t\u00eb mbeten t\u00eb dallueshme.<\/li><\/ul>

                                                  M\u00ebsimi 25 \u2013 Dam\u00eb turke (Dama):<\/h1>

                                                  Alfa-Beta, MCTS i udh\u00ebhequr nga PUCT, Alpha Zero<\/h1>

                                                  Zbaton dam\u00ebn turke dhe krahasohet me k\u00ebrkimin klasik (alpha\u2013beta) me MCTS duke p\u00ebrdorur nj\u00eb ekzekutues ndeshjesh t\u00eb rip\u00ebrdorsh\u00ebm dhe log-e simulimi n\u00eb grup.<\/p>

                                                  • Motor loje: p\u00ebrfaq\u00ebsimi i bordit, l\u00ebvizjet e lejuara me kapje me shum\u00eb k\u00ebrcime dhe kodimi i rrug\u00ebs s\u00eb l\u00ebvizjes.<\/li>
                                                  • Funksioni i vler\u00ebsimit plus agjenti i k\u00ebrkimit Negamax\/Alpha-Beta; kompromiset midis thell\u00ebsis\u00eb dhe forc\u00ebs.<\/li>
                                                  • Agjenti MCTS p\u00ebr damat turke dhe krahasimet ball\u00eb p\u00ebr ball\u00eb kund\u00ebr alfa\u2013beta.<\/li>
                                                  • Universal match runner (play_game) dhe utilitetet e simulimit n\u00eb grup p\u00ebr eksperimente t\u00eb riprodhueshme.<\/li>
                                                  • Regjistrime t\u00eb eksportueshme (t\u00eb kompresuara n\u00eb zip) p\u00ebr rishikim dhe debugim n\u00eb klas\u00eb.<\/li><\/ul>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                                                    \n\t\t\t\t
                                                    \n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>","protected":false},"excerpt":{"rendered":"

                                                    Shatranj.ai project curriculum can be accessed through lms.shatranj.ai Below are the short summaries of topics within the curriculum. Lesson 1 \u2013 Course Scope and Priorities Introduces the Shatranj.AI project, its Erasmus+ foundations, partner organizations, and digital platforms. Project vision, Erasmus KA2 context Partner institutions and cultural heritage focus Overview of platforms (editor, LMS, code tools) […]<\/p>\n","protected":false},"author":11,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-195","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/shatranj.ai\/sq\/wp-json\/wp\/v2\/pages\/195","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/shatranj.ai\/sq\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/shatranj.ai\/sq\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/shatranj.ai\/sq\/wp-json\/wp\/v2\/users\/11"}],"replies":[{"embeddable":true,"href":"https:\/\/shatranj.ai\/sq\/wp-json\/wp\/v2\/comments?post=195"}],"version-history":[{"count":12,"href":"https:\/\/shatranj.ai\/sq\/wp-json\/wp\/v2\/pages\/195\/revisions"}],"predecessor-version":[{"id":820,"href":"https:\/\/shatranj.ai\/sq\/wp-json\/wp\/v2\/pages\/195\/revisions\/820"}],"wp:attachment":[{"href":"https:\/\/shatranj.ai\/sq\/wp-json\/wp\/v2\/media?parent=195"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}