Reinforcement Learning on Atari 57 (Individual Game Scores)

3,824,506.3AtlantisLBC

Updated 5mo ago

Evaluation Results

Method
LBC2023.05	3,824,506.3	279,703.5	12,996.3	62,025.7	999,999	1,106,603.5	1,410	857,369	457,321	35,340	233.1	100	864	728,080	999,999	233,090	995,950	900,170	24	14,332.5	75	34	13,792.4	488,900	6,372.5	37,545.6	47.53	623,300.5	14,372.6	593,679.5	1,666,665	2,500	31,403	81,473	999,999	-1	21	15,100	151,730	27,964.3	999,999	144	1,000,000	-5,903.34	10,732.5	159,999.6	999,999	2.726	24	531,614	436.2	999,999	2,200	999,999	118,900	998,970	241,570.6	—	—	10,077.52	1,665.6	—	—
MuZero2020.03	1,674,767.2	741,812.63	28,634.39	143,972.03	998,425	6,785,558.64	1,278.98	848,623	4,549,993.53	85,932.6	260.13	100	864	1,159,049.27	991,039.7	458,315.4	839,642.95	143,964.26	23.94	2,382.44	91.16	33.03	631,378.53	130,345.58	6,682.7	49,244.11	67.04	41,063.25	16,763.6	269,358.27	204,824	0	243,401.1	157,177.85	955,137.84	0	21	15,299.98	72,276	323,417.18	613,411.8	131.13	999,976.52	-29,968.36	56.62	74,335.3	549,271.7	9.99	0	476,763.9	491.48	715,545.61	0.4	981,791.88	197,126	553,311.46	725,853.9	—	—	—	—	—	—
MEME2023.05	1,669,226.33	83,683.43	14,368.9	46,635.86	769,803.92	364,492.07	87,792.55	776,770	51,870.2	38,838.35	261.74	99.85	831.08	245,892.18	912,225	339,274.67	543,979.5	142,176.58	23.7	2,360.64	77.05	33.97	526,239.5	119,457.53	20,875	199,880.6	47.22	117,009.92	17,311.17	155,915.32	476,539.53	12,437	29,747.91	40,077.73	849,969.25	46,734.79	19.31	100,798.9	238,453.5	90,333.12	399,511.83	114.46	960,181.39	-3,273.43	28,175.53	57,828.45	264,286.33	9.82	22.79	404,751.67	1,030.27	524,631	2,859.83	617,640.95	71,942	633,867.66	77,942.17	—	—	4,081.14	1,225.19	—	—
R2D22020.03	1,660,721.85	464,232.43	31,331.37	110,100.04	999,354.03	431,072.45	27,117.85	992,600.31	390,603.06	77,725.62	161.77	100	863.92	908,137.24	999,900	729,482.83	730,714.53	143,913.32	24	2,378.66	90.34	34	309,077.3	129,736.13	21,068.03	49,339.62	86.59	158,142.36	18,284.99	245,315.44	267,766.63	3,000	62,595.9	138,030.67	990,638.12	0	21	40,700	777,071.3	93,569.66	593,186.78	144	999,999	-3,851.44	67,306.29	67,898.71	998,600.28	10	24	460,596.49	483.78	702,700.36	2,258.93	999,645.92	183,090.81	999,807.02	370,649.03	—	—	—	—	—	—
Agent572020.03	1,528,841.76	297,638.17	29,660.08	67,212.67	991,384.42	150,854.61	23,071.5	934,134.88	300,509.8	61,507.83	251.18	100	790.4	412,847.86	999,900	565,909.85	677,642.78	143,161.44	23.93	2,367.71	86.97	32.59	541,280.88	117,777.08	19,213.96	114,736.26	63.64	135,784.96	24,034.16	251,997.31	206,845.82	9,352.01	63,994.44	54,386.77	908,264.15	18,756.01	20.67	79,716.46	580,328.14	63,318.67	243,025.8	127.32	999,997.63	-4,202.6	44,199.93	48,680.86	839,573.53	9.5	23.84	405,425.31	2,354.91	623,805.73	2,623.71	992,340.74	157,306.41	998,532.37	249,808.9	—	—	—	—	—	—
Agent572023.05	1,528,841.76	297,638.17	29,660.08	67,212.67	991,384.42	150,854.61	23,071.5	934,134.88	300,509.8	61,507.83	251.18	100	790.4	412,847.86	999,900	565,909.85	677,642.78	143,161.44	23.93	2,367.71	86.97	32.59	541,280.88	117,777.08	19,213.96	114,736.26	63.64	135,784.96	24,034.16	251,997.31	206,845.82	9,352.01	63,994.44	54,386.77	908,264.15	18,756.01	20.67	79,716.46	580,328.14	63,318.67	243,025.8	127.32	999,997.63	-4,202.6	44,199.93	48,680.86	839,573.53	9.5	23.84	405,425.31	2,354.91	623,805.73	2,623.71	992,340.74	157,306.41	998,532.37	249,808.9	—	—	4,762.17	1,933.49	—	—
ASL DDQN2023.05	947,275	6,955.2	2,232.3	14,372.8	567,640	1,984.5	1,340.9	38,986	26,841.6	2,597.2	62.4	99.6	621.7	3,899.8	15,071	166,019	37,026.5	119,773.9	0.1	2,103.1	35.1	33.9	8,616.4	103,514.4	760	26,578.5	-3.6	2,237	13,027	10,422.5	85,182	0	4,416	16,535.4	71,752.6	0	21	349.7	24,548.8	24,445	56,520	65.8	29,278.6	-8,295.4	3,506.8	21,602	129,140	2.5	22.3	12,071	252.9	25,127.4	291	626,794	21,049	29,231.9	16,420	—	—	—	—	—	—
Noisy DQN2023.05	923,733	2,403	1,610	5,510	14,328	3,455	1,068	36,786	20,793	905	71	89	516	4,269	8,893	118,305	20,525	36,150	0.1	1,240	11	32	753	14,574	447	6,246	—	1,235	10,944	8,805	36,310	3	2,722	8,181	16,028	0	21	3,712	15,545	9,425	45,993	51	2,282	-14,763	6,088	2,186	47,133	-1	0	7,035	232	14,255	97	322,507	9,198	23,915	6,920	—	—	—	—	—	—
MuZero + Past Policies2021.06	845,409	52,821	4,239	35,013	549,421	235,543	552	72,183	130,129	35,249	47	99	770	534,432	660,503	217,204	483,394	112,140	-1	132	-52	0	8,049	120,551	3,927	9,871	-11	13,668	10,465	11,295	83,705	0	43,700	94,974	49,919	-24	-6	-6	70,593	28,026	571,829	25	141,725	-30,062	1,501	5,367	547,226	6	0	228,282	150	432,240	0	378,897	54,093	423,271	15,790	—	—	—	—	—	—
MuZero (our impl.)2021.06	835,445	38,698	6,631	35,876	674,573	214,034	837	39,471	120,675	22,449	59	99	504	400,268	524,655	189,621	322,472	131,963	3	0	-97	0	3,439	121,984	2,807	7,877	-6	23,475	9,659	11,259	55,242	0	40,263	76,604	67,119	-2	-7	193	64,732	27,688	151,639	53	27,530	-27,968	1,544	3,962	663,896	7	-23	267,331	134	434,746	0	376,660	79,425	317,803	15,752	—	—	—	—	—	—
Ape-X DDQN2023.05	831,952	2,596.1	1,271.2	14,490.1	189,300	1,107.9	1,264.9	38,671	27,033.7	924.8	60.2	99.4	370.7	3,808.4	6,031	118,020	29,255	114,874.7	-0.2	1,969.1	31.2	21.4	504.7	47,845.6	242.5	14,464	-2.5	540	14,710	10,999.4	54,124	0	4,087.7	16,042.7	28,296	0	21	173	15,300	22,238	51,208	42.6	32,101.8	-15,623.8	1,523.2	3,943	60,835	2.9	—	7,457	226.2	46,208	62	603,075	14,780	13,178.5	—	—	—	—	—	—	—
DSQN2023.05	487,366.7	—	—	—	—	—	—	—	7,226.9	—	—	95.3	386.5	—	—	123,916.7	—	—	1	—	—	—	—	10,107.3	—	—	—	1,156.7	8,880	9,940	—	—	—	10,877	—	—	20.3	—	—	—	48,983.3	—	—	—	—	1,832.2	57,686.7	—	—	—	194.7	—	—	275,342.8	—	—	—	—	—	—	—	—	—
Duel. DDQN2023.05	382,572	4,461.4	2,354.5	4,621	28,188	2,837.7	1,611.9	37,150	12,164	1,472.6	65.5	99.4	345.3	7,561.4	11,215	143,570	42,214	60,813.3	18.5	2,258.2	46.4	0	4,672.8	15,718.4	588	20,818.2	0.5	1,312.5	14,854	11,451.9	34,294	0	6,283.5	11,971.1	23,092.2	0	21	103	19,220.3	21,162.6	69,524	65.3	50,254.2	—	2,250.8	6,427.3	89,238	4.4	5.1	11,666	211.4	44,939.6	497	98,209.5	7,855	49,622.1	12,944	—	—	—	—	—	—
Prior. DDQN2023.05	357,324	4,203.8	1,838.9	7,672.1	31,527	2,654.3	1,054.6	31,530	23,384.2	1,305.6	47.9	95.6	373.9	4,463.2	8,600	141,161	31,286.5	71,846.4	-6.3	2,093	39.5	33.7	4,380.1	32,487.2	548.5	23,037.7	1.3	5,148	16,200	9,728	39,581	0	6,518.7	12,270.5	18,992.7	-356.5	20.6	200	16,256.5	14,522.3	57,608	62.6	26,357.8	-9,996.9	4,309	2,865.8	63,302	8.9	0	9,197	204.6	16,154.1	54	282,007.3	4,802	11,357	10,469	—	—	—	—	—	—
DQN2023.05	85,641	3,069	739.5	3,359	6,012	1,629	429.7	26,300	6,846	—	42.4	71.8	401.2	8,309	6,687	14,103	—	9,711	—	301.8	-0.8	30.3	328.3	8,520	306.7	19,950	-1.6	576.7	6,740	3,805	23,270	0	2,311	7,257	—	—	18.9	1,788	10,596	8,316	18,257	51.6	5,286	—	—	1,976	57,997	—	-2.5	5,947	186.7	8,456	380	42,684	3,393	—	4,977	—	—	—	—	—	—
DDQN2023.05	64,758	2,907.3	702.1	5,022.9	15,150	930.6	728.3	25,730	7,654	—	70.5	81.7	375	4,139.4	4,653	101,874	—	9,711.9	-18.1	319.5	20.3	31.8	241.5	8,215.4	—	20,357	-2.4	438	13,651	4,396.7	29,486	0	3,210	6,997.1	—	—	21	670.1	14,875	12,015.3	48,377	46.7	7,995	—	—	3,154.6	65,188	—	1.7	7,964	190.6	16,769.9	93	70,009	5,204	—	10,182	—	—	—	—	—	—
Average Human2020.03	29,028.1	7,127.7	1,719.5	742	8,503.3	47,388.7	753.1	37,187.5	16,926.5	2,630.4	160.7	12.1	30.5	12,017	7,387.8	35,829.4	18,688.9	1,971	-16.4	860.5	-38.7	29.6	4,334.7	2,412.5	3,351.4	30,826.4	0.9	302.8	3,035	2,665.5	22,736.3	4,753.3	6,951.6	8,049	7,242.6	6,463.7	14.6	69,571.3	13,455	17,118	7,845	11.9	42,054.7	-4,336.9	12,326.7	1,668.7	10,250	6.5	-8.3	5,229.2	167.6	11,693.2	1,187.5	17,667.9	4,756.5	54,576.9	9,173.3	—	—	—	—	—	—
Average Human2023.05	29,028.1	7,127.8	1,719.5	742	8,503.3	47,388.7	753.1	37,187.5	16,926.5	2,630.4	160.7	12.1	30.5	12,017	7,387.8	36,829.4	18,688.9	1,971	-16.4	860.5	-38.8	29.6	4,334.7	2,412.5	3,351.4	30,826.4	0.9	302.8	3,035	2,665.5	22,736.3	4,753.3	6,951.6	8,049	7,242.6	6,463.7	14.6	69,571.3	13,455	17,118	7,845	11.9	42,054.7	-4,336.9	12,326.7	1,668.7	10,250	6.5	-8.3	5,229.2	167.6	11,693.2	1,187.5	17,667.9	4,756.5	54,576.9	9,173.3	—	—	—	—	—	—
Random2020.03	12,850	227.8	5.8	222.4	210	719.1	14.2	2,360	363.9	123.7	23.1	0.1	1.7	2,090.9	811	10,780.5	2,874.5	152.1	-18.6	0	-91.7	0	65.2	257.6	173	1,027	-11.2	29	52	1,598	258.5	0	307.3	2,292.3	761.4	-229.4	-20.7	24.9	163.9	1,338.5	11.5	2.2	68.4	-17,098.1	1,236.3	148	664	-10	-23.8	3,568	11.4	533.4	0	0	563.5	3,092.9	32.5	—	—	—	—	—	—
RND2023.05	12,850	227.8	5.8	222.4	210	719	14.2	236	363.9	123.7	23.1	0.1	1.7	2,090.9	811	10,780.5	2,874.5	152.1	-18.6	0	-91.7	0	65.2	257.6	173	1,027	-11.2	29	52	1,598	258.5	0	307.3	2,292.3	761.5	-229.4	-20.7	24.9	163.9	1,338.5	11.5	2.2	68.4	-17,098	1,236.3	148	664	-10	-23.8	3,568	11.4	533.4	0	0	563.5	3,092.9	32.5	—	—	—	—	—	—
Random2023.05	12,850	227.8	5.8	222.4	210	719.1	14.2	2,360	363.9	123.7	23.1	0.1	1.7	2,090.9	811	10,780.5	2,874.5	152.1	-18.6	0	-91.7	0	65.2	257.6	173	1,027	-11.2	29	52	1,598	258.5	0	307.3	2,292.3	761.4	-229.4	-20.7	24.9	163.9	1,338.5	11.5	2.2	68.4	-17,098.1	1,236.3	148	664	-10	-23.8	3,568	11.4	533.4	0	16,256.9	563.5	3,092.9	32.5	—	—	—	—	—	—